Clinical Characteristics in Patients with Hospital-Acquired Pneumonia Caused by Extensively Drug-Resistant Acinetobacter baumannii, A Multicenter Retrospective Study

Background Extremely drug-resistant (XDR) Acinetobacter baumannii (cid:0) A. baumannii (cid:0) has been of a great concern. The relationship between XDR and patient outcomes remains unclear. We investigated the clinical features, risk factors, and outcomes of Hospital-acquired pneumonia (HAP)caused by XDR A. baumannii. Methods A multicenter retrospective case-control study was performed to determine factors associated with XDR A. baumannii pneumonia from 5 teaching hospitals in Guangzhou, China. Results 76 patients were enrolled in the study. XDR A. baumannii pneumonia patients were tend to be smoker (11.9% vs 3.9%, P = 0.130) and older (76.5±11.2 vs 70.3±16.4, P = 0.007) and had more comorbid diseases including chronic obstructive pulmonary disease (COPD) (48.7% vs 21.1%, P = 0.001) and renal failure (21.1% vs 3.9%, P = 0.002) and had higher APACHE II score (65.8% vs 47.4%, P = 0.033). Invasive procedures including insertion of urinary catheter, nasogastric tube, central venous/arterial catheter, bronchoscopy and mechanical ventilation along with using β-lactam/β-lactamase inhibitor and carbapenem were also risk factors for XDR A. baumannii pneumonia. Multivariate analysis showed the APACHE II score >=20 (OR, 2.1; 95% CI: 1.1–4.1, P = 0.023), COPD (OR, 9.6; 95% CI: 2.0–45.5, P = 0.004), central venous/arterial catheter placement (OR,11.5; 95% CI: 1.1-117.8, P = 0.040), low albumin levels (OR, 1.2; 95% CI: 1.1-1.4, P = 0.001) and using β-lactam/β-lactamase inhibitor (OR,15.9;

Conclusions XDR A. baumannii pneumonia was strongly related to systemic illnesses, invasive procedure, low albumin levels and the APACHE II score and increasing the length of mechanical ventilation and hospital stay. But it did not increase in-hospital mortality.

Background
Acinetobacter baumannii (A. baumannii) is one of the most troublesome pathogens causing Hospitalacquired pneumonia (HAP) worldwide [1][2][3]. Due to its remarkable ability to acquire antibiotic resistance, A. baumannii infection is also di cult to control and treat [4][5][6]. CHINET surveillance system from China showed A. baumannii resists to many powerful antimicrobial agents, especially imipenem and meropenem which were once considered as the "gold standard" [7]. And the resistance rate to those agents went up to 77.1% and 78.1% respectively in 2018 [6]. Extremely drug-resistant (XDR) A. baumannii was de ned as resistance to all available antibiotics except for colistin and tigecycline [8][9][10]. It became a major resistance phenotype, ranged from 68.3-72.9% of the reported A. baumannii strains [11,12]. Colistin and tigecycline are believed to be the last few available choices for XDR infections. But tigecycline and polymyxin resistance A. baumannii had already been reported [5,[13][14][15]. XDR A. baumannii with limited treatment options was believed to be associated with high mortality and treatment failure and served as a marker of poor prognosis.
However, the relationship between antibiotic resistance of A. baumannii strains and outcomes caused with concern and dispute. One study in 2007 found mortality rates of patients with multi-drug resistant (MDR) A. baumannii were not higher than the patients without MDR A. baumannii [16]. Our previous study in 2017 was limited by the sample size, but it showed XDR A. baumannii pneumonia was not associated with in-hospital mortality [10]. Two studies in 2019 showed increasing mortality rate may relate to patient comorbidity and inappropriate therapy, but not antibiotic resistance [12,17]. It is frustrated for clinicians to treat XDR A. baumannii and the data for XDR A. baumannii HAP was still limited. We believe the relationship between XDR A. baumannii pneumonia and outcomes needs further investigation. Here, we conducted a multicenter study to investigate the clinical features, risk factors, and outcomes of HAP caused by XDR A. baumannii from 5 teaching hospitals in Guangzhou, China.  [12] from XDR A. baumannii (con rmed with respiratory samples) from April 2011 to April 2016. The controls, matched to the cases by hospitals, were patients diagnosed with non-XDR A. baumannii HAP and randomly selected within the 5 teaching hospitals during the studying period.

Inclusion and exclusion criteria
The inclusion criteria consisted of the following: 1, Diagnostic criteria for pneumonia [18]: new or progressive pulmonary in ltrates in chest X-ray or CT imaging, plus at least two of the following supportive clinical signs: temperature of > 38 °C or < 35.5 °C, leukocytosis of > 12,000 or < 4000 WBC/mm 3 , purulent respiratory secretions, or worsening oxygenation determined by arterial oxygen tension/fraction of inspired oxygen ratio (PaO 2 /FiO 2 ).
3, Antimicrobial susceptibility testing shows resistance to all available antibiotics except for colistin and tigecycline.
The exclusion criteria consisted of the following: 1, Patient was less than 18 years of age.
2, The length of hospitalization was less than 48 hours.

Data acquisition
A specially designed case record form (CRF) was used to collect the patient's clinical parameters, including age, gender, cigarette smoking, substance use history, laboratory test results (white blood cell count, platelet count, hemoglobin, creatinine, CRP, ESR and albumin), past medical diseases, invasive procedures (venous or arterial catheter placement, bronchoscopy, mechanical ventilation, et al.), the length of hospital stay (including ICU), pharmaceutical records (including antibiotic exposure), microbiological records (bacterial etiology, culture, and antimicrobial susceptibility testing results) and inhospital mortality. The Acute Physiology and Chronic Health Evaluation (APACHE) II score was used to determine the severity of diseases on patient's admission. The in-hospital mortality was compared between the HAP with and without XDR A. baumannii infection groups.

Results
Based on the inclusion and exclusion criteria, a total of 76 patients with XDR A. baumannii pneumonia were included in the case-control study. 53 were male and 23 were female. The mean age was 76.5-yearold with a standard deviation of 11.2 years (Table 1). To analyze the potential risk factors, 76 cases of patients were compared with 76 control patients.  Discussion XDR A. baumannii infection rate kept rising. But the antibiotics treatment options were very limited. It made the XDR A. baumannii one of the most troublesome pathogens worldwide. It seems preventing spreading the pathogen is the only reasonable way to reduce mortality. However, recently, the relationship between antibiotic resistance and outcomes caused with concern and dispute. Here, we presented this multicenter retrospective case-control study to investigate the clinical features, risk factors, and outcomes of HAP caused by XDR A. baumannii. It showed XDR A. baumannii pneumonia signi cantly increased the length of mechanical ventilation and hospital stay. But it was not associated with in-hospital mortality. Our nding indicated, besides antibiotics, we could improve outcomes if we optimize treatment in XDR A. baumannii pneumonia patients. This study may provide an insight view of XDR A. baumannii pneumonia clinical management during an era of stagnation in new antibiotic development.
Our study found that XDR A. baumannii pneumonia patients were more likely to have a smoking history and in elderly patients with complicated underlying disease and severer initial presentation ( higher APACHE II score). It was similar to previous studies [19][20][21]. Multivariate logistic regression analysis showed the APACHE II score greater than 20, COPD, and lower albumin levels were the independent risk factors for XDR A. baumannii pneumonia. Those ndings indicated immunocompromised patients were more likely to develop XDR A. baumannii HAP. Using β-lactam/β-lactamase inhibitors was also another independent risk factor. Clinicians need to be extra careful while using β-lactam/β-lactamase inhibitors.
XDR A. baumannii pneumonia patients may receive more invasive procedures due to severer clinical presentation. But medical devices may be the most contaminated locations for A. baumannii and caused hospital-acquired infections outbreaks [22,23]. Therefore, clinicians should be aware of invasive procedures uses and removing unnecessary accesses in time.
Previous studies had shown patients with antibiotic-resistant A. baumannii may have increased mortality compared with antibiotic-sensitive patients. [16,24,25]. The study in 2018 showed the increased mortality rate was associated with patient comorbidity and inappropriate therapy, but not antibiotic resistance [26]. And more studies revealed antibiotic resistance of A. baumannii was not associated with increase mortality. Sunenshine RH, et al [16] reported mortality rate of MDR A. baumannii infection patients was not higher than non-MDR A. baumannii patients. Özgür et al [9] also found the mortality rate of patients with XDR A. baumannii VAP was not signi cantly higher than non-XDR A. baumannii patients in ICU settings. Aušra Cˇ iginskiene, et al [12] reported the mortality rate in XDR A. baumannii VAP patients was not a signi cant difference compared with the non-XDR group. Zilberberg et al [17]

Declarations
Ethics approval and consent to participate The Guangzhou First People's Hospital Ethics Committee approved the study. Informed consents were not required as this was considered a review of clinical practice.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
YjL supervised the study and wrote the manuscript. XmH, WyY and CzP discussed the data and helped nalize the manuscript. ZxM, ZxZ, McH, CqF, HlC and PhG collected the data. ZwZ and SqW planned and supervised the experiments. All authors read and approved the nal manuscript.