Clinical and Mortality Risk Factors in Bloodstream Infections with Carbapenem-Resistant Enterobacteriaceae

Objective To investigate the risk factors underlying the occurrence and mortality of bloodstream infections (BSIs) with carbapenem-resistant Enterobacteriaceae (CRE). Methods Medical information was retrospectively analyzed from 148 cases of patients with Enterobacteriaceae BSIs at a medical center in China, between 2013 and 2015. Results The 30-day mortality rate in the CRE group was 65.4%. Indwelling urethral catheterization, admission to the ICU, use of antibiotics within 30 days, and BSIs from the respiratory system were associated with CRE BSIs. Lung infection, abdominal infection, central venous catheterization, and use of hormones within 30 days were associated with mortality. Conclusion The 30-day mortality rate of CRE BSIs was high. Lung infections, abdominal infections, central venous catheterization, and use of hormones within 30 days increased the mortality rate of Enterobacteriaceae BSIs.


Introduction
Recently, infections with drug-resistant bacterial strains have increased which poses a challenge for anti-infection treatments in the clinical context and has become a serious problem in the eld of public health. e incidence of bloodstream infections is also increasing and has become a major cause of the occurrence of infectious diseases and deaths worldwide [1]. Bloodstream infections with multidrugresistant Enterobacteriaceae account for 71.5% of all bloodstream infections with multidrug-resistant bacterial strains [2]. Carbapenem antibiotics are typically one of the most e ective drug classes for the treatment of Enterobacteriaceae infections. However, carbapenem-resistant Enterobacteriaceae (CRE) has been identi ed recently within all Enterobacteriaceae spp. e CHINET monitoring results show that Klebsiella pneumoniae accounts for the majority of cases of CRE infections among the Enterobacteriaceae spp., with drug resistance rates to imipenem and meropenem > 10% [3]. Studies have shown that the mortality rate is signi cantly higher for CRE infections than for carbapenem-sensitive Enterobacteriaceae infections [4].
is study was a retrospective analysis of medical information on bloodstream infections with Enterobacteriaceae (mainly Klebsiella pneumoniae, Escherichia coli, Enterobacter cloacae, and Enterobacter aerogenes) with intent to identify the risk factors and prognosis for CRE bloodstream infections.

Study Subjects.
A total of 148 cases of bloodstream infections with Enterobacteriaceae that occurred during the period between January 2013 and October 2015 in Fuxing Hospital of Capital Medical University, China, were selected. Clinical information from the patients, including demographic data, underlying diseases, pathogen drug resistance and drug sensitivity analyses, history of medical institution admission, level of care (ward versus ICU), history of invasive surgery, and prognosis, was recorded and analyzed. e characteristics of each case were collected from the electronic medical record database, and the les of all patients were complete.
Inclusion criteria: On the day of collection of a positive blood culture specimen, the patient had a temperature ≥ 38°C or < 36°C in combination with chills, and the multiple blood culture results showed the same type of bacteria. Exclusion criteria: (1) blood culture test results showing two or more types of bacteria and (2) only one blood culture specimen was positive, whereas multiple subsequent blood culture results were negative or identi ed other types of pathogens.
is study conformed to the standards of medical ethics and was approved by the Ethics Committee of Fuxing Hospital of Capital Medical University.

Relevant De nitions.
e microbiology laboratory in our hospital currently uses an FX200 automatic blood culture instrument (BACTE) and a WalkAway 40 automatic microbiological drug sensitivity analysis instrument (USA). e drug sensitivity results were determined according to the Clinical and Laboratory Standards Institute (CLSI, USA) standards [5]. e de nition of CRE was resistance to imipenem, meropenem, or ertapenem [3]. Extensively drugresistant (XDR) bacteria were de ned based on nonsusceptibility to all antimicrobial agents except colistin and tigecycline [6,7]. Long-term bedbound patients were dened as patients who were bedbound for more than 14 days and could not recover [7].

Statistical Methods.
e data processing and analysis were performed using SPSS 21.0 statistical software. P < 0.05 indicated that the di erence was signi cant. Measurement data with a normal distribution are expressed using the mean ± standard deviation. Measurement data with a nonnormal distribution are expressed using the median. Significance testing for di erences between groups was performed using the t-test (normal distribution) or rank sum test (nonnormal distribution). Signi cance testing for di erences in count data between groups was performed using the χ 2 test. e multivariate analysis was performed using logistic regression.

Demographic Data and Clinical Characteristics.
is study involved a total of 148 cases of bloodstream infections with Enterobacteriaceae. e mean age of the patients was 69.1 (±19.2) years. Male patients accounted for 50.7% of the cases, and 20.3% of the patients were admitted to the ICU. Long-term bedbound patients accounted for 54.1% of the cases, and 62.2% of the patients had a medical institution admission history within 3 months. e bloodstream infections mainly originated from the urinary tract (28.4%) or respiratory system (25.0%) or were catheter related (22.3%). Patients who had prior use of broad-spectrum antibiotics within 30 days accounted for 43.2% of the cases. e total average length of hospital stay was 21.7 (±12.5) days. Patients with CRE infections accounted for 17.6% of the cases, and patients with XDR Enterobacteriaceae infections accounted for 4.0% of the cases. e results of the univariate and multivariate analyses of risk factors related to CRE bloodstream infections are shown in Tables 1 and 2, respectively (the relevant factors listed in the tables were all conditions before the presence of the bloodstream infection). e logistic regression analysis revealed that indwelling urethral catheterization, admission to the ICU, CRE infection of the respiratory system, and prior use of broad-spectrum antibiotics within 30 days were independent risk factors for CRE bloodstream infections.

Discussion
Currently, the incidence of carbapenemase-producing Enterobacteriaceae is increasing. Data have shown that many countries have interregional spread or an endemic situation [8]. e results of this study showed that the 30-day mortality rate for CRE bloodstream infections reached 65.4%, which was signi cantly higher than the rate for non-CRE bloodstream infections (17.2%) or for CRE-infected patients in literature reports (26%-44%) [4]. To date, many studies have evaluated risk factors related to bloodstream infections with multidrug-resistant Enterobacteriaceae. For example, factors including admission to the ICU, length of hospital stay, history of use of broad-spectrum antibiotics, such as quinolones and cephalosporins, history of resistant strain colonization, indwelling urethral catheterization, and central venous catheterization have been considered independent risk factors for bloodstream infections with multidrug-resistant Enterobacteriaceae [4,9]. e analytical results in this study showed that indwelling urethral catheterization, admission to the ICU, prior use of broadspectrum antibiotics within 30 days, and CRE infection of the respiratory system were independent risk factors for CRE bloodstream infections. e rst three factors are similar to previous reports. However, what is di erent from previous reports is that we identi ed CRE infection of the respiratory system as an independent risk factor for CRE bloodstream infections. Respiratory tract infection was the most common source of bloodstream infection in our study, likely because of a decreased capacity for bacterial clearance resulting from the altered lung tissue in severe chronic obstructive lung disease. e results of this study on risk factors for death due to bloodstream infections with Enterobacteriaceae showed that lung infections, abdominal infections, central venous catheterization, and hormone use within 30 days were all independent risk factors for death due to bloodstream infections with Enterobacteriaceae. e study results showed that the mortality rate was signi cantly higher in the CRE group than in the non-CRE group. e high mortality  associated with CRE bloodstream infections was partially attributed to host conditions, such as endogenous infections and invasive surgery. Additionally, the mortality rate was associated with a lack of timely and accurate antibiotic treatment for infection with drug-resistant bacterial strains or even no antibiotic treatment [10]. XDR strains accounted for 23% of the 26 cases of CRE infections in this study, which was similar to the 28% rate of XDR bacterial strains isolated from  CRE infections reported in the literature [8]. is result indicated that approximately one-fourth of CRE infections might be XDR, which is associated with a high mortality rate. During the analysis of mortality-related risk factors in patients with bloodstream Enterobacteriaceae infections in this study, CRE infection was a relevant but not an independent risk factor, which might be associated with the low overall number of cases.
In summary, CRE bloodstream infections have a high mortality rate, and treatment has poor clinical e cacy. e use of broad-spectrum antibiotics, indwelling urethral catheterization, admission to the ICU, and infection from the respiratory system were independent risk factors for CRE bloodstream infections. Lung infections, abdominal infections, hormone use, and central venous catheterization were independent risk factors for death in patients with Enterobacteriaceae bloodstream infections. erefore, the standard use of antibiotics, minimization of invasive operations, use of strict aseptic manipulation, and strengthening of handhygiene awareness of healthcare workers may be e ective measures for the prevention of CRE bloodstream infections.

Conflicts of Interest
e authors declare that there are no con icts of interest.