Nosocomial Infections Due to Multidrug-Resistant Bacteria in Cancer Patients: A Six-Year Experience of an Oncology Center of Western China

Background: Bacterial infections are the most frequent complications in patients with malignancy, and the epidemiology of nosocomial infections among cancer patients has changed over time. This study aimed to evaluate characteristics, antibiotic-resistant patterns, and prognosis of nosocomial infections caused by multidrug-resistant (MDR) bacteria in cancer patients. Methods: This retrospective observational study analyzed cancer patients with MDR bacteria caused nosocomial infections from August 2013 to May 2019. The extracted clinical data were recorded in a standardized form and compared based on the patient’s survival status after infection during hospitalization. Data were analyzed by using independent samples t-test, Chi-square test, and binary logistic regression. P -values < 0.05 were considered statistically significant. Results: Overall, 257 cancer patients developed nosocomial infections caused by MDR bacteria. Extended-spectrum β-lactamase producing Enterobacteriaceae (ESBL-PE) was the most frequently isolated multidrug-resistant Gram-negative bacteria (MDRGNB), followed by Acinetobacter baumannii , and Stenotrophomonas maltophilia . Smoking history, cancer patients who received intrapleural/abdominal infusion within 30 days, presence of indwelling urinary catheter, length of hospitalization, and haemoglobin were independent factors for in-hospital mortality in the study population. The isolated MDR bacteria were mainly sensitive to amikacin, meropenem, imipenem, tigecycline, and piperacillin/tazobactam. Conclusions: Cancer patients with prolonged hospitalization was an independent predictor of a favorable outcome. However, former smokers, cancer patients who received intrapleural/abdominal infusion within 30 days, presence of indwelling urinary catheter, and anemia were independent risk factors of in-hospital mortality. Our findings suggest that clinicians should think highly of nosocomial infections caused by MDR in cancer patients and advise policymakers to develop a guideline.

suggested that infection in cancer patients is associated delayed initiation of chemotherapy, reduced standard dosage, prolonged hospitalization, increased financial burden of healthcare, and raised severe morbidity and mortality [3][4][5][6]. The epidemiology of nosocomial infections among cancer patients has changed over time, and the causative organisms of nosocomial infections had shifted from Gram-positive pathogens to Gram-negative pathogens in the last 20 years worldwide [7][8][9].
Most of the previous studies [10][11][12][13] showed that extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE), multidrug-resistant (MDR) Pseudomonas aeruginosa, carbapenemresistant Enterobacteriaceae (CRE), Acinetobacter baumannii and methicillin-resistant Staphylococcus aureus (MRSA) have been increasingly identified as the predominant causative infection pathogens in cancer patients due to the phenomenon of antibiotics misuse [14,15]. Previously published guidelines recommend antibiotic treatment for cancer patients with neutropenia or septic shock, while there were no clearly defined empirical antibiotic regimens for highly suspected MDR bacteria caused nosocomial infection episodes in cancer patients [16][17][18]. Therefore, therapeutic options for these infections are often limited, and the prognoses are generally worse for these patients. We conducted this present six-year retrospective study to demonstrate the clinical characteristics, microbial spectrum, antibiotic resistance patterns, and prognostic factors for nosocomial infections in cancer patients caused by MDR bacteria between August 2013 and June 2019 at the First Affiliated Hospital of Xi'an Jiaotong University.

Study population and design
We conducted a single-center retrospective observational study in a 2560-bed university referral cancer center in Xi'an, China. The study enrolled cancer patients who received medical care during hospitalization from August 2013 to May 2019. The electronic medical record database was reviewed to identify cancer patients with nosocomial infections caused by MDR bacteria. Patients with hematological malignancies were all excluded, and only the initial infection episode was analyzed.

Data collection
Data in electronic medical records of all included cancer patients were extracted. The extracted clinical data included age, gender, smoking history, Eastern Cooperative Oncology Group (ECOG) performance status, primary location of the disease, existence of distant metastasis, American Joint Committee on Cancer (AJCC) TNM categories, primary sites of infection, comorbidities and severity of underlying conditions according to the Charlson comorbidity index (CCI) [19], existence of fever, types of cancer therapy within 30 days (surgery, chemotherapy, radiotherapy, or concurrent chemoradiotherapy), corticosteroid treatment within previous 30 days, prior infection before hospital admission, Granulocyte colony-stimulating factor (G-CSF) use within 30 days, previous antibiotics treatment within 30 (7) MRSA [20].
MDR strains were defined as resistance to at least one agent in each of three or more categories of antimicrobial agents, including β-lactam/β-lactamase inhibitor combinations (piperacillin/tazobactam), extended-spectrum cephalosporins (ceftriaxone, ceftazidime, cefepime), carbapenems (imipenem/meropenem), monobactams, aminoglycosides (gentamicin, amikacin) and/or fluoroquinolones, and if they were resistant to vancomycin for Enterococcus faecium, and resistant to methicillin for Staphylococcus aureus [16]. Clinical samples such as sputum, urine, blood culture, stool, wounds secreta, ascites, pleural, drainage fluid post-operation, and other samples were collected once patients were suspected of MDR infection. The swabs were collected for colonization screening but not for the diagnosis of MDR infection episodes.
Antimicrobial susceptibility of isolated organisms was determined using the Kirby-Bauer disk diffusion method according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [21]. ESBL resistant organisms were defined as resistant to one or more extended-spectrum cephalosporins, and ESBL producers were confirmed via the double-disk synergy test.
Nosocomial infection was defined as signs or symptoms of infection that occurred >48 hours after hospital admission or <48 hours after hospital discharge. Otherwise, the case was considered community-onset [22].
Fever was considered as an axillary temperature of 38.3 °C on one occasion or a temperature of >38.0 °C on two or more occasions during 12 hours [23].
Empirical antibiotics treatment was considered as the initiation of antimicrobial agents before the results of microbiology and communicated to clinicians [15]. Empirical antibiotics treatment was considered effective once the antibiotics used could suppress the activity of the isolated MDR pathogens according to the results of antimicrobial susceptibility tests [15].

Study outcomes
This study aimed to describe the clinical characteristics, microbial spectrum, antibiotic resistance patterns, and prognostic factors of all cancer patients with MDR bacteria caused nosocomial infections and to present in-hospital mortality and its associated risk factors. In-hospital mortality was defined as death during hospitalization for the studied infection episodes.

Statistical analysis
The extracted clinical data were recorded in a standardized form and compared based on the patient's survival status after infection during hospitalization. Parametric continuous quantitative variables were described as means and standard deviation, while median and interquartile ranges were used for non-parametric continuous variables. Continuous variables were analyzed by the independent samples t-test or the Mann-Whitney U test. Categorical variables were analyzed by the chi-square or Fisher's exact tests. Univariate and multivariate logistic regression analyses were used to investigate risk factors for in-hospital mortality of cancer patients with nosocomial infections caused by MDR bacteria. Variables with a p-value of <0.10 from univariate analysis and variables with clinical significance were included in a multivariate logistic regression analysis using stepwise selection. All statistical analyses were performed by the SPSS software version 22.0 for Windows (SPSS Inc., Chicago, IL, USA).  Table 1).

Infection-related data of cancer patients with nosocomial infections caused by MDR bacteria
We reviewed all of the clinical data of MDR bacteria caused nosocomial infections in cancer patients.

Comparison of clinical and infection-related characteristics in the study population based on the patient's survival status during hospitalization
We used in-hospital mortality to evaluate the primary clinical outcomes of MDR bacteria caused nosocomial infections in cancer patients. Among the study subjects, the overall case-fatality rate was 10.9% (28/257). We also analyzed the relationship between prognosis and clinical characteristics of these infections in cancer patients. The results showed that smoking history, ECOG performance status, existence of distant metastasis, presence of liver disease, CCI, existence of fever, underwent surgery or chemotherapy within 30 days, received intrapleural/abdominal infusion within 30 days, presence of indwelling catheters or other devices (indwelling urinary catheters and drains postoperation) significantly differed between survivors and non-survivors (P<0.05; Table 1). Additionally, sample type (sputum and urine), primary sites of infection (respiratory tract infection, urinary tract infection, and BSI), length of hospitalization, septic shock, and laboratory examination results (haemoglobin, platelet count, lymphocytes count, and albumin) have also differed between survivors and non-survivors (P<0.05; Table 2).

Bacterial characteristics
During the seven years, there were 257 cultures isolated from different clinical specimens. The great majority of specimens were urine (36.2%), followed by sputum (24.1%), blood culture (17.1%), wounds secreta (8.2%), and drainage fluid post-operation (8.2%). The annual distribution showed that the MDR bacteria caused nosocomial infection episodes mainly distributed from 2014 to 2018, despite only partial data were collected in 2013 and 2019 (Figure 2a). Besides, it also showed that the majority of patients had MDR bacteria caused nosocomial infections that were detected with a peak during August and September of each year (Figure 2b). The causative pathogens of nosocomial infection episodes are compared by survival status during hospitalization in Table 3. Overall, ESBL-PE was the most frequently isolated multidrug-resistant Gram-negative bacteria (MDRGNB) (72.4%), followed by Acinetobacter baumannii (11.7%), Stenotrophomonas maltophilia (6.2%), MDR Pseudomonas aeruginosa (5.1%), Carbapenem-resistant Enterobacteriaceae (1.6%), and ESBLproducing Enterobacter aerogenes (0.4%). MRSA was the only isolated multidrug-resistant Grampositive bacteria, accounting for 2.7%. However, there was no significant difference in causative pathogen distribution (P>0.05) between the survivor and non-survivor groups ( Table 3).

Risk factors for in-hospital mortality among cancer patients with nosocomial infections caused by MDR bacteria
In this study, the univariate analysis demonstrated that smoking history, ECOG performance status,  (Table 1 & 2).

Antimicrobial susceptibility analysis
The antimicrobial sensitivity of commonly used antibiotics showed that the isolated MDRGNB were mainly sensitive to piperacillin/tazobactam, meropenem, imipenem, amikacin, and tigecycline, while they were mainly resistant to ceftriaxone, aztreonam, and ciprofloxacin. MRSA was the only isolated MDR Gram-positive bacteria, and the drug sensitivity analysis showed that these strains were mainly sensitive to vancomycin, linezolid, moxifloxacin, levofloxacin, and tigecycline (Figure 3).

Discussion
In this retrospective study, we found the prevalence of MDR bacteria caused nosocomial infections in cancer patients was 25.5%, which was consistent with the recent study conducted in Spain (25.5%) [24]. However, in the Spain study, the subjects were cancer patients with FN, and all of the organisms were isolated from blood culture. At the same time, a prospective observational study conducted in In this study, we found that ESBL-PE was the most frequent organism caused nosocomial infection episodes in cancer patients, accounting for 72.4%, followed by Acinetobacter baumannii (11.7%), and Stenotrophomonas maltophilia (6.2%). Compared with Gram-negative bacteria, MRSA was the only isolated MDR bacteria in Gram-positive bacteria, accounting for 2.7%. Biehl LM et al. [25] reported that ESBL-producing Enterobacteriaceae is emerging as a new threat to both nosocomial and community infections worldwide, and ESBL-PE caused approximately 1 in 10 nosocomial infection episodes in patients with malignancy. At the same time, ESBL-PE caused nosocomial infections was more worrisome due to the increased mortality in these patients [25]. Therefore, rapid initiation of appropriate and adequate antibiotic therapy is pivotal for nosocomial infection episodes caused by ESBL-PE, and since most empirical regimens do not adequately cover these pathogens [1].
Among the study subjects, a considerable overall case-fatality rate of 10.9% was observed in our study, which was lower compared with most of the previous studies conducted in other countries [15,[26][27][28][29]. According to a retrospective conducted in Brazil, Freire MP et al. reported that the overall case-fatality rate of carbapenem-resistant K. pneumoniae caused nosocomial infection reached 57.8% in patients with solid tumor as well despite its small sample size in this cohort (83 infection episodes) [26]. In a case-control study, including 204 cancer patients admitted to ICU, Nazer LH et al. reported that 30-day mortality exceed 70% for cancer patients with Acinetobacter baumannii caused nosocomial infection [28]. At the same time, Moghnieh R et al. reported that the case-fatality rates of MDR bacteria caused nosocomial infection up to 57.1% in cancer patients with FN [29]. Recently, in a five years period retrospective study including 73 patients with solid tumors, Perdikouri EIA et al.
reported that 30% of cases were died due to MDR bacteria-caused infection [15]. This may have been due to the majority of patients in this study who were at an advanced stage and have distant metastasis.
The results of the multivariate analysis identified that former smokers were associated with a higher case-fatality rate in cancer patients with MDR bacteria caused nosocomial infections, which was an interesting finding in our study. Stämpfli MR et al. reported that cigarette exposure significantly impacts the immune system, impairing the host's ability to produce appropriate immune and inflammatory responses and promoting infection [30]. We found that cancer patients who received intrapleural/abdominal infusion within 30 days were associated with a higher case-fatality rate in our cases, possibly suggesting that these patients are more likely to exposure MDR bacterial infection, catheter-related infection, and can be easily immunocompromised [2]. The presence of indwelling urinary catheters was also found to be an independent risk factor for mortality in cancer patients with MDR bacteria caused nosocomial infections. This finding is consistent with previous studies conducted elsewhere in patients with MDR caused nosocomial infections [31][32][33]. Our study demonstrated that prolonged hospitalization was an independent factor for a favorable outcome in these patients. On the contrary, Perdikouri EIA et al. reported that prolonged hospitalization was associated with an increased fatality rate in cancer patients [15]. On the one hand, it may be attributed to the majority of cases in this study suffered from metastasis disease, and received invasive procedures during hospitalization. On the other hand, we used in-hospital mortality to evaluate the clinical outcomes of infection episodes in our study, and the median length of hospitalization was 21.0 days in our study.
Therefore, it needs to be further validated before drawn conclusions. Besides, anemia was also found to be as independent risk factors for mortality in cancer patients with MDR bacteria caused nosocomial infections. Zhang LN et al. reported that pretreatment anemia-induced tissue hypoxia might directly reduce the overall survival of cancer patients [34].
The antimicrobial susceptibility showed that the isolated MDRGNB were mainly sensitive to piperacillin/tazobactam, meropenem, imipenem, amikacin, and tigecycline, while they were mainly resistant to aztreonam, cephalosporins (third or fourth generation), and fluoroquinolone. MRSA was the only isolated MDR Gram-positive bacteria, and the drug sensitivity analysis showed that these strains were mainly sensitive to vancomycin, linezolid, moxifloxacin, levofloxacin, and tigecycline, which was comparable with previous studies [35,36]. The phenomenon of MDR can be attributed to the overuse of antibiotics in China. There is an urgent need to execute and implement policies on the control of antibiotics misuse to avoid the evolution of newer generations of highly resistant pathogens. Besides, the entire microbial spectrum should be taken into consideration when initiating empirical antibiotic treatment [35]. As far as we know, this is the first study evaluated clinical characteristics, microbial spectrum, antibiotic resistance patterns, and prognostic factors among cancer patients with MDR bacteria caused nosocomial infections in China, and one of the advantages of our study is that we have studied the vital number of risk factors. However, our study has several limitations. First, it is hard to collect some variables (e.g., concrete chemotherapeutic or radiation dosage, concrete antibiotics treatment before admission, and some laboratory examination results) in this retrospective study. Thus, there might be hidden biases in the analysis of the relationship. In addition, our study was a single-center retrospective study. Therefore, a prospective multicenter study is needed for further validation.

Conclusions
In conclusion, the overall case-fatality rate in cancer patients with MDR bacteria caused nosocomial infections was 10.9%. The most frequently isolated pathogens were ESBL-PE, Acinetobacter baumannii, and Stenotrophomonas maltophilia. Prolonged hospitalization was an independent predictor of a favorable outcome. However, former smokers, Cancer patients who received intrapleural/abdominal infusion within 30 days, presence of indwelling urinary catheters, and anemia were independent risk factors for in-hospital mortality in the study subjects. The isolated MDR bacteria were mainly sensitive to piperacillin/tazobactam, meropenem, imipenem, amikacin, and tigecycline. We suggest that clinicians should think highly of nosocomial infections caused by MDR in cancer patients, and consider the epidemiological characteristics of local resistant patterns when initiating antimicrobial treatment.

Ethics approval and consent to participate
The study was approved by the ethics committee of the First Affiliated Hospital of Xi'an Jiaotong University. Waiving of informed consent was obtained due to the retrospective noninterventional study design.

Consent for publication
Not applicable.

Availability of data and material
Please contact author for data requests.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author contributions
TT and YY conceived the study. AMJ and XS were involved in data collecting, statistical analysis, and drafting the manuscript. NL, HG, and MDR carried out the data collection and analysis and provided the critical revision. XQZ and XF participated in the study design and helped with the data collection.