Study setting and design
A prospective observational study carried out in Zhongshan Hospital, Fudan University (Shanghai, China), an 2430-bed tertiary institution(15 Respiratory ICU beds and 22 Emergency ICU beds), where located at east of China covering 10 million of Shanghai’s population and radiating 150 million people in the surrounding of Yangtze River Delta. The medical system and accessibility to medical care is no different to the other parts of China. All consecutive cases of severe CAP visiting the emergency department or respiratory clinic from August 2017 to December 2018 were included. The study was approved by the Institutional Review Board at Zhongshan Hospital, Fudan University Human Subjects Research Protection Program Office (B2017–061R). All patients were written the informed consents.
One thousand and three hundred seventeen patients had symptoms suggestive of lower respiratory infection were referred to the emergency department as generally accepted. 806 patients were diagnosis of pneumonia. The diagnosis for CAP was 695 patients with a new chest radiographic infiltrate on computed tomography (CT) scan , symptoms including cough and/or sputum productive, fever or hypothermia, and abnormal leukocytes counts [1, 6]. 127 patients were admitted to ICUs according to pulmonologist’s judgement. The objective criteria of severe CAP was according to 2007 IDSA/ATS guidelines . The CURB–65 criteria (confusion, uremia, respiratory rate, low blood pressure, age 65 years or greater) , and prognostic models, such as the Pneumonia Severity Index (PSI) , were also reference. The study profile was shown in Figure 1.
All patients with the exacerbation of chronic respiratory disease (CRD, included chronic obstructive pulmonary disease (COPD), asthma, bronchiectasis and interstitial pneumonia, which have acute worsening of respiratory symptoms to additional therapy without new chest radiographic infiltrate), healthcare associated pneumonia (HCAP, pneumonia in non-ambulatory residents of nursing homes and other long-term care facilities) [1,19], hospital-acquired pneumonia (HAP, any pneumonia contracted by patients in hospital at least 48–72 hours after being admitted) , severe immune-compromised (acquired immune deficiency syndrome, organ transplant, and tumor chemotherapy)  and age less than 18 years were excluded.
Data collection and follow up
The clinical information of all patients were carefully collection including age, sex, smoking history (smoking ≥ one cigarette per day for one year), alcohol abuse (consumption≥120 g per day), comorbidity, initial symptoms, the CURB scores and PSI severity, acute physiology and chronic health evaluation-II (APACHEII), sequential organ failure assessment (SOFA) , biochemistry and arterial blood gas measurements, chest radiograph features, mechanical ventilation requirement, length of hospital stay, mortality etc. Microbiological evaluation was collected emphasis on diagnostic methods, pathogens distribution, and MDR status. The follow up visit was performed in 30–90 days after discharge from ICUs.
At least two blood sample and sputum specimens collected within 24 hours after admission were investigated. If the patient’s temperature was more than 38.5℃, blood culture was needed. The qualified sputum specimens were screened by microscopy (if < 10 epithelial cells and > 25 leukocytes per field; magnification 100) . The methods of urine antigen were detected S. pneumoniae and Legionella pneumophila. Nasopharyngeal swabs for respiratory virus and serology for atypical microorganisms were also detection. Paired serology was analyzed by means of at admission and during the third week thereafter for Mycoplasma pneumoniae, Chlamydophila pneumonia. Legionella pneumophila and respiratory virus were diagnosed using the indirect immunofluorescence technique for antibodies detection. Pleural fluid culture was performed for patients with significant pleural effusion on chest radiograph. Bronchoalveolar lavage (BAL) and lung tissue were performed by fiberoptic bronchoscopy .
All samples were cultured in adequate media that allowed optimal results. The etiology of pneumonia was classified as presumptive if a valid sputum sample yielded one or more predominant bacterial strain; The etiology was considered by blood culture yielding a bacterial or fungal pathogen; High titers of immunoglobulin (Ig)M antibodies in the serum during the acute phase was accepted or four-fold increase in IgG titers for the diagnosis of atypical microorganisms .
Statistical analyses for clinical characteristics and outcomes were reported as means±standard deviation or as percentages. The continuous variables were statistically analyzed by a one-way analysis of variance test. The categorical variables were assessed with chi-square test with Fisher’s exact test correction. Calculations of odds ratio (OR) and 95% CI values for severe CAP in relation to potential risk factors on mortality were performed with binary logistic regression models ; these covariates included age, gender, smoking history, alcohol abuse, comorbidities, APACHE II, SOFA, the most three frequent pathogens (K. pneumoniae, Acinetobacter baumannii (A. baumannii)and S. aureus) and mixed infections.
All tests were two-sided, and a P value of 0.05 was considered statistically significant. Statistical analyses were conducted with IBM SPSS for windows, version 24.0.