In this observational cohort study, we retrospectively reviewed the medical records of patients with pneumonia admitted to an ICU via ER at a teaching hospital between March 2015 and May 2017. We excluded patients who were transferred from other acute care hospitals or had received treatment for pneumonia within the previous 3 months.

Pneumonia was diagnosed based the presence of a new opacity development and at least two of the following four clinical criteria: fever or hypothermia (body temperature >38℃ or ≤35℃), leukocytosis or leukopenia (white blood cell count ≥10,000/µL or ≤4,000/µL), newly developed respiratory symptoms (cough, sputum, pleuritic chest pain, and dyspnea), and altered breath sounds on auscultation15. We defined NHAP as pneumonia occurring in a resident of a nursing home or a long-term care facility1. All patients were admitted to an ICU for close monitoring or after requiring mechanical ventilation with shock or acute respiratory failure in an ER.

The Institutional Review Board of Inje University Ilsan Paik Hospital approved the study design, including the review and publishing of information obtained from patient records (IRB No. 2017-12-018). The requirement for informed consent was waived for the use of patient medical data because all personally identifying information was removed before analysis.

Patient medical records were reviewed to obtain data on clinical characteristics, clinical parameters, laboratory findings, clinical outcomes, and isolated microorganisms. Comorbidities including diabetes mellitus, degenerative nerve disease, chronic obstructive pulmonary disease, malignant neoplasm, and chronic renal disease were reviewed. Clinical parameters included the CURB-65 (confusion, urea, respiratory rate, blood pressure, age more than 65 years) for assessing the severity of pneumonia16, use of vasopressors, and need of mechanical ventilator. Prior antibiotics use within 90 days was arbitrarily defined as use of antibiotics within 90 days for reasons other than pneumonia.

Microbiological studies were conducted using two sets of blood cultures, gram staining and culture using sputum, tracheal aspirate, and bronchial washing fluids that were obtained by bronchoscopy. Respiratory specimens were cultured in a semi-quantitative manner, and pathogens were identified when a predominant microorganism was detected from group 4 or 5 sputum, according to Murray and Washington's grading system. Blood cultures were considered to be positive if pathogens were present, there were no other infection sources that could explain a positive culture, and the possibility of contamination was excluded. Urinary antigen test for Streptococcus pneumoniae was also considered to indicate etiological pathogens. The antibiotic sensitivity of all isolated pathogens was identified using a disc diffusion method. Multidrug-resistant gram-negative bacteria (MDRGNB) for Pseudomonas aeruginosa, Enterococcus species, Enterobacteriaceae, and Acinetobacter species was defined as acquired non-susceptibility to at least one agent in three or more antimicrobial categories17.

All patients were initially administered broad empirical antibiotics according to the ATS/IDSA guideline18. The detailed antibiotic regimen complied with the attending physician's choice, taking into consideration any risk factors of the patient. In this study, the use of inappropriate antibiotics was defined as the use of empirical antibiotics which were not effective or which were unnecessarily broad against the identified pathogens based on in vitro susceptibility testing6.

The data are presented as medians and interquartile ranges for continuous variables and as numbers (%) for categorical variables. The data were compared using the Mann-Whitney U test for continuous variables and Pearson's chi-square test or Fisher exact test for categorical variables. Multivariate logistic regression analysis was used to determine independent factors associated with the use of inappropriate antibiotics in patients with HCAP. All tests were 2-sided and a p-value <0.05 was considered significant. All statistical analyses were performed using IBM SPSS Statistics for Windows version 22.0 (IBM Corp., Armonk, NY, USA).

From the medical records of 172 patients with pneumonia admitted to an ICU via ER, after excluding 28 patients who were transferred from other acute care hospitals and 22 patients who had received treatment for pneumonia within the previous 3 months, a total of 122 patients with CAP and NHAP were included. Among 122 patients, 83 with identified causative bacteria were included in the present study; 39 patients were excluded because a causative agent could not be identified. Of the 83 patients, 29 patients had CAP (34.9%) and 54 had NHAP (65.1%) (Figure 1). In this study, 50 patients (63.9%) and 30 patients (36.1%) received appropriate and inappropriate antibiotics, respectively. Among the total 88 patients, respiratory viral real-time polymerase chain reaction was performed in 48 (57.8%) and 16 (19.3%) patients were confirmed to have bacterial and viral co-infections. The clinical characteristics of the study subjects are summarized in Table 1. NHAP patients were more frequently treated with inappropriate antibiotics rather than appropriate antibiotics (47.2% [25/53] vs. 96.7% [29/30], p<0.001). A CURB-65 score ≥3 was more common in the inappropriate antibiotic-treated group than in the appropriate antibiotic-treated group (25.8% [31/53] vs. 70.0% [21/30], p=0.046). There were no significant differences between the clinical parameters, laboratory parameters, and identified microorganisms between the two groups. The differences in clinical outcomes, including duration of overall admission, duration of ICU admission, and 30-day mortality, between the two groups were not significant.

In 29 CAP patients and 54 NHAP patients admitted to the ICU via ER, the majority of them received combination antibiotic therapy as the initial treatment (CAP 71.4%, NHAP 66.7%) (Table 2). Among the 54 NHAP patients, 25 (46.3%) received appropriate antibiotics. Patients treated with combination antibiotics that included vancomycin were more frequently encountered in the appropriate antibiotics group than in the inappropriate antibiotics group (20.0% [5/25] vs. 0% [0/29], p=0.017).

The identified microorganisms in the study subjects are summarized in Table 3. Streptococcus pneumoniae was more frequently isolated from subjects in the appropriate antibiotics group than from those in the inappropriate antibiotics group (20.8% [11/53] vs. 0% [0/30], p<0.001). Staphylococcus aureus was more frequently isolated from subjects in the inappropriate antibiotics group than from those in the appropriate antibiotics group (28.3% [15/53] vs. 70% [21/30], p<0.001). The prevalence of the other identified microorganisms was not significantly different between the two groups. Methicillinresistant Staphylococcus aureus (MRSA) was more frequently isolated from subjects in the inappropriate antibiotics group than from those in the appropriate antibiotics group (7.5% [4/54] vs. 70% [21/30], p<0.001). The prevalence of MDRGNB was not significantly different between the two groups.

Candidate variables for logistic regression analysis included prior antibiotics use within 90 days, NHAP, identification of polymicrobial pathogens, and MDRGNB, which were objective variables with p<0.1 when the appropriate antibiotics group and inappropriate antibiotics group were compared (Table 4). In multivariate analysis, identification of NHAP (adjusted odds ratio, 28.66; 95% confidence interval, 3.45??38.12; p=0.002) was independently associated with the use of inappropriate antibiotics.