Clinical risk factors and outcomes of carbapenem-resistant Escherichia coli nosocomial infections in a Chinese teaching hospital: a retrospective study from 2013 to 2020

ABSTRACT The emergence of carbapenem-resistant Escherichia coli strains poses a considerable challenge to global public health, and little is known about carbapenemase-producing E. coli strains in Tianjin, China. This study aimed to investigate the risk factors for infections with carbapenem-resistant E. coli (CREC) strains. This retrospective case–control study was conducted at a tertiary teaching hospital. A total of 134 CREC clinical isolates were collected from the General Hospital of Tianjin Medical University between 2013 and 2020. The control group was selected at a ratio of 1:1 from patients with nosocomial carbapenem-susceptible E. coli infection. Risk factors for nosocomial CREC infection and clinical outcomes were analyzed using univariate and multivariate analyses. Multivariate analysis revealed that cephalosporin exposure (odd ratio OR = 2.01), carbapenem exposure (OR = 1.96), glucocorticoid exposure (OR = 32.45), and surgical history (OR = 3.26) were independent risk factors for CREC infection. The in-hospital mortality rate in the CREC group was 29.1%, and age >65 years (OR = 3.19), carbapenem exposure (OR = 3.54), and central venous catheter insertion (OR = 4.19) were independent risk factors for in-hospital mortality in patients with CREC infections. Several factors were identified in the development of nosocomial CREC infections. The CREC isolates were resistant to most antibiotics. Reducing CREC mortality requires a comprehensive consideration of appropriate antibiotic use, underlying diseases, and invasive procedures. IMPORTANCE Escherichia coli is an opportunistic pathogen that causes severe hospital-acquired infections. The spread of carbapenem-resistant E. coli is a global threat to public health, and only a few antibiotics are effective against these infections. Consequently, these infections are usually associated with poor prognosis and high mortality. Therefore, understanding the risk factors associated with the causes and outcomes of these infections is crucial to reduce their incidence and initiate appropriate therapies. In our study, several factors were found to be involved in nosocomial carbapenem-resistant E. coli (CREC) infections, and CREC isolates were resistant to most antibiotics. Reducing CREC mortality needs a comprehensive consideration of whether antibiotics are used appropriately, underlying diseases, and invasive interventions. These findings provide valuable evidence for the development of anti-infective therapy, infection prevention, and control of CREC-positive infections.

The CSEC group was selected from the same ward as the source population during the same period as the CREC group (within 2 weeks); if two or more controls were matched, one was randomly selected.

Micro-biologic methods
All isolates were identified using matrix-assisted laser desorption ionization time-of-flight mass spectrometry (bioMerieux, France).An automated broth microdilution method (VItek2; bioMerieux) was used for susceptibility testing.The Clinical and Laboratory Standards Institute (CLSI-M100) document was used to interpret the antimicrobial susceptibility testing (23).E. coli ATCC25922 and Pseudomonas aeruginosa ATCC27853 were used as control strains for antimicrobial susceptibility testing.

Data collection
We reviewed the medical records and collected patient information (20)(21)(22).The epidemiology and clinical data of patients were collected, including the department, sex, age, age >65 years, underlying diseases (respiratory system, liver, urinary system, circulatory system, and central nervous diseases; sepsis; digestive system diseases; diabetes mellitus; malignancy; and immunosuppression), antibiotics (third-generation cephalosporin, β-lactam inhibitor compounds, carbapenem, quinolones, tigecycline, and macrolides); glucocorticoids, within 3 months prior to CREC infections; surgical history and invasive procedures (mechanical ventilation, central venous catheter, arterial and urinary catheters, drainage, and gastric tubes) within 1 month prior to a positive culture; and mortality, related to hospitalization (hospital stays prior to CREC isolation, intensive care unit (ICU) stays within 3 months prior to a positive culture, and total duration of hospital stays).
Patients with a positive culture from blood or any other sterile source were defined as having an infection, and patients with positive cultures from respiratory, urine, and surgical wounds were defined as having an infection based on the Centers for Disease Control and Prevention (CDC) and National Healthcare Safety Network (NHSN) criteria (24).
The clinical outcomes were defined as follows: 30-day mortality (within 30 days of the first positive culture), in-hospital mortality (death during hospitalization after the first culture), and total duration of hospital stay (duration from admission to hospital-to-hos pital discharge).

Statistical analysis
Normally distributed continuous variables are presented as mean ± standard deviation (SD) and were compared using a t-test.Non-normally distributed continuous variables are presented as medians with interquartile ranges (IQRs) and were compared using the Mann-Whitney U-test.Categorical variables are presented as counts and percentages and were compared using the χ test or Fisher's exact test.
Univariate analyses were performed for each variable, and those with a P < 0.05 were included in a multivariate logistic regression analysis.Statistical significance was set at P < 0.05, and SPSS 26.0 was used for statistical analyses.

Specimen source ward and source site
A total of 134 patients were included in this study.Most patients infected with CREC were admitted to the ICU (47%, 63/134), and then in general medical wards (36.6%, 49/134), and surgical wards (16.4%, 22/134).The most common types of infections were respiratory tract, urinary tract, and bloodstream infections (Table 1).

Resistance rate to antibiotics
The antibiotic susceptibility patterns of CREC and CSEC isolates are shown in Table 2.All the CREC isolates were resistant to ampicillin, cefazolin, cefuroxime, ceftriaxone, and ciprofloxacin.The resistance rates of CREC to amoxicillin/clavulanic acid, cefoper azone/sulbactam, piperacillin /tazobactam, ceftazidime, cefepime, aztreonam, levofloxacin, and trimethoprim/sulfamethoxazole were higher than 80%, whereas the resistance rate to amikacin was relatively low.

DISCUSSION
To the best of our knowledge, few studies have evaluated the risk factors for the acquisition of CREC infection, with most reports focusing on the mechanisms of carbapenem resistance in CRE (25)(26)(27)(28).Therefore, the aim of this matched case-control study was to assess the potential risk factors for CREC infections.
In this study, we found that CREC infections were mainly concentrated in the ICU, followed by the emergency department in the general medical ward and the  neurosurgical department in the surgical ward.Patients in the ICU have serious underlying diseases and require life-support systems and broad-spectrum antibiotics (29).Patients with various infectious diseases are often treated in the emergency department.Patients undergoing craniocerebral surgery need to have a drainage tube in place for a long time and use broad-spectrum antibiotics because of potential intracra nial infections.
All the CREC strains were resistant to ampicillin, cefazolin, cefuroxime, ceftriaxone, and ciprofloxacin.Antibiotic resistance was severer in the CREC group than in the CSEC group, and the differences between the antibiotics listed in Table 2 in the two groups were statistically significant.However, these strains were relatively susceptible to amikacin.Considering the aforementioned results and the individual clinical conditions, the treatment of CREC infections was optimal.Our results indicate that exposure to antibiotics (third-generation cephalosporin and carbapenems), glucocorticoids, and surgical history are risk factors for CREC infection.
Studies have reported a close association between CREC infection and antibiotic use, particularly carbapenem exposure (30)(31)(32)(33).Our study showed that prior carbapenem use was an independent risk factor for CREC infection, which is consistent with the results of most previous studies.Acquisition of the ability to produce carbapenemase is the main cause for carbapenem resistance, which can easily result in nosocomial spread (34).Tian et al. demonstrated that the most prevalent carbapenemase gene in CREC in China is blaNDM, followed by blaKPC-2 (10).However, Dagher et al. found that the predominant carbapenemase detected was OXA-48 in Lebanon (35), and that carbapenem exposure may induce the emergence of these resistance-conferring genes.In our study, third-generation cephalosporin exposure within 3 months was also a risk factor for CREC infection, indicating that CREC infection can be induced not only by the use of carbapenem but also by the use of other drugs.Thus, we need to strengthen the  management of antibiotics for inpatients, and treatment with high doses for controlled durations is a better way to limit the risk of infection (32).
Unlike other studies, our study revealed that previous use of glucocorticoids was also a risk factor for CREC infection, possibly because patients with CREC infections have serious underlying diseases and are treated with glucocorticoids.Second, the use of glucocorticoids destroys the intestinal microenvironment, kills sensitive strains, and promotes the overgrowth of drug-resistant strains, which promotes the shift of opportunistic pathogens to pathogenic bacteria.It is not surprising that surgical history was a risk factor for CREC infection, which is consistent with previous studies emphasizing the importance of aseptic operations in patient care (36).The asepsis technique is an important strategy for preventing CREC infections.
In our study, gastric tube insertion was identified as an independent risk factor for 30-day mortality.Meta-analyses revealed that the use of medical devices, including gastric tubes, presents the highest aggregated estimate (37).Furthermore, multiple studies have proposed that CREC colonization of the gastrointestinal tract contributes to subsequent CREC infection in afflicted individuals (38).Additionally, a significant proportion of patients undergoing gastric tube insertion exhibited gastrointestinal dysfunction and a compromised nutritional status, which could facilitate bacterial translocation in the gastrointestinal tract and potentially fostering pathogenic bacterial infections, including CREC.CREC infection and carbapenem exposure were independ ent risk factors for 30-day mortality.Therefore, the judicious use of carbapenems in antibiotic stewardship programs, reduction of invasive procedures, and prevention of CREC infections are effective ways to reduce 30-day mortality.
Patients with CREC infections who died were significantly older than surviving patients with CSEC infection.Possible reasons for this include more frequent healthcare exposure and antibiotic exposure among older adults (39).Central venous catheter insertion was an independent risk factor for in-hospital mortality in patients with nosocomial CREC infections.The most likely reason for this is that invasive procedures can damage the mucosa and increase the incidence of CREC infection because the majority of the bacteria are able to pass through the mucosal barrier to other sites (40).Carbapenem exposure was also a risk factor for in-hospital mortality in patients with CREC infection.The possible reasons for this are the same as mentioned previously.
Recently, an increasing number of studies have focused on active screening for CRE, enabling early detection, isolation, and intervention in high-risk departments or patients.The World Health Organization (WHO) and the US Centers for Disease Control (CDC) recommend screening for CRE based on local epidemiological conditions (41,42).In our hospital, nosocomial CREC infections tend to occur in specific departments and in patients with risk factors, suggesting that those higher-risk departments should actively screen for CREC.
This study has several limitations that should be considered.First, a molecular epidemiological investigation of CREC was not performed; thus, we could not assess whether there were any outbreaks during the study period and whether different drug resistance mechanisms resulted in different clinical outcomes; future studies should include molecular epidemiological investigations.In addition, analyzing risk factors through case-control studies has the potential to increase the utilization of antibiot ics, which can introduce statistical selection bias.This is due to the fact that certain antibiotics may inhibit or eliminate certain strains of CSEC, while having no effect on CREC.Consequently, the frequency of antibiotic usage in CSEC strains may decrease, indirectly amplifying the risk factor for CREC (43).Therefore, attention should be paid to the selection of the control group, as choosing an antibiotic-sensitive group as the control group may increase the risk of antibiotic exposure (44).Notably, this was a single-center study with a modest sample size.Hence, it is imperative to conduct prospective, multicenter, large-sample clinical trials to validate our findings.

Conclusion
The aim of this study of nosocomial CREC infection was to establish a foundation for the formulation of efficacious control strategies.In our hospital, CREC infections exhibited a focal distribution across specific departments, such as the ICU, emergency depart ment, and neurosurgery department, underscoring the need to implement interven tions such as hand hygiene protocols and environmental disinfection measures to limit CREC transmission.Paying close attention to patients who have used carbapenem and cephalosporins for a long time, the rational use of antibiotics, and the reduction of invasive procedures are effective measures to reduce CREC infection and 30-day or in-hospital mortality.These findings provide valuable evidence for the development of anti-infective therapy, infection prevention, and control of CREC-positive infections.

FIG 3
FIG 3 Kaplan-Meier curves showing 30-day mortality in the CREC group versus the CSEC group (P = 0.009).

TABLE 2
The antibiotic resistance of CREC and CSEC groups a a CREC, carbapenem-resistant Escherichia coli; CSEC, carbapenem-sensitive Escherichia coli.

TABLE 3
Demographic characteristics of the CREC and CSEC groups a a SD, standard deviation; CREC, carbapenem-resistant Escherichia coli; CSEC, carbapenem-sensitive Escherichia coli; LOS, length of hospital stays; IQR, interquartile range.

TABLE 4
Univariate analyses regarding the risk factors for CREC infections a a CREC, carbapenem-resistant Escherichia coli; CSEC, carbapenem-sensitive Escherichia coli.TABLE 5 Multivariate analyses regarding the risk factors for CREC infections

TABLE 6
Clinical outcome comparison between the CREC group and the CSEC group a a CREC, carbapenem-resistant Escherichia coli; CSEC, carbapenem-sensitive Escherichia coli; LOS, length of hospital stays; IQR, interquartile range.

TABLE 7
Univariate analyses of in-hospital mortality in the CREC group a

TABLE 8
SD, standard deviation; CREC, carbapenem-resistant Escherichia coli; ICU, intensive care unit; LOS, length of hospital stays; IQR, interquartile range; -, no result.Multivariate analyses of in-hospital mortality in the CREC group a a a CREC, carbapenem-resistant Escherichia coli.