Comparison of the results of sputum culture and bronchoscopic lavage fluid culture:an analysis report of large sample size data from the Clinical Microbiology Laboratory

Background Sputum specimens were the most common for Clinical Microbiology Laboratory in China. For the results of sputum culture, it was difficult for clinicians to evaluate the significance, and even more difficult for laboratory physicians. At present, most Clinical Microbiology Laboratories in China executed quality assessment of sputum specimens. But how to evaluate the results of sputum culture was still very confusing. To solve this problem, we conducted a series of retrospective studies. Methods Based on the culture results of bronchoscopic lavage fluid (BALF), the differences of sputum culture results before (2013-2015) and after (2016-2018) quality control of sputum samples in our hospital were compared. Results Pseudomonas aeruginosa , Acinetobacter baumannii , Klebsiella pneumoniae and Staphylococcus aureus were the four most common pathogens in sputum and BALF culture, both in 2013-2015 years and 2016-2018 years. Antimicrobial susceptibility test from 2013-2015 and 2016-2018 both showed for P. aeruginosa and K. pneumoniae , the susceptibility rates of BALF isolates to all commonly used antibiotics were higher than those from sputum. For A. baumannii and S. aureus , the sensitivity rates of BALF isolates to most antibiotics were higher than those from sputum. After quality control of sputum samples, there was still a difference between the results of sputum culture and those of BALF. Conclusions Even though the quality control of sputum specimens had been carried out, the results of culture and antimicrobial resistance of pathogens from qualified sputum samples were still different from those of BALF.

3 Background Lower respiratory tract infections (LRIs) were the most common infectious disease of respiratory tract. [1][2] Pneumonia was the second most common infection in hospitalized patients, and was highly correlated with morbidity and mortality. [3] Irrational use of antibiotics would delay the patient's condition and cause serious bacterial resistance. [1][2] According to the data from China Antimicrobial Resistance Surveillance System (CARSS) in 2015, the major specimens type from inpatients in respiratory departments in China were sputum (81. 6%, 41,131/50,417). [1][2] Due to the convenience of specimens collection, sputum had always been the most common type of specimens in clinical microbiology laboratories in China. But easily confused by oral colonization flora, it was difficult to judge the results of sputum culture as infectious, colonized or contaminated bacteria. For the results of sputum culture, it was difficult for clinicians to evaluate the significance, and even more difficult for laboratory physicians.
It was well known that unqualified sputum specimens, such as saliva, have no significance in bacterial culture. And the industry had reached a consensus that sputum culture without microscopic examination was of no value. [4] As early as the 1970s, the quality control of sputum specimens had been studied. Generally, sputum specimens were judged to be qualified by the number of inflammatory cells (primary polymorphonuclear leukocytes) and epithelial cells examined by a stained smear of the specimen. According to six different interpretation criteria from

Study design and procedures
A retrospective analysis was made of the differences between sputum culture results and BALF culture results during the period of no quality control of sputum specimens from 2013 to 2015 and the period of quality control of sputum specimens from 2016 to 2018 in our hospital. For the main pathogenic bacteria, the difference of antimicrobial sensitivity between sputum specimens and BALF was analyzed.

Source of specimens
All specimens (Including sputum specimens and BALF specimens) were taken from the clinical departments of Tongji Hospital and sent to department of laboratory medicine.
Interpretation criteria for qualified sputum specimens 5 The qualified interpretation of sputum specimens in this study is based on the Chinese standard.
[6] Since 2016, smear microscopy has been required for each specimen requiring sputum culture. Sputum specimens satisfying the following three conditions will be treated as qualified specimens. First, a specimen with ≥25 white blood cells (WBC) per average low-power field (LPF) and squamous epithelial cells (EPI) <25 per LPF. Second, the ratio of WBC to EPI was more than 10:1, and the single-form bacteria were predominant. Third, EPI <10 per LPF, and alveolar macrophages and columnar epithelial cells existed. In addition to the above three cases, when EPI> 10 per LPF, they were considered to be unqualified sputum specimens.

Identification of strains and antimicrobial susceptibility test
For isolates from the same position of the same patient, only the first isolate was included in the analysis according to CLSI M39.
Antimicrobial susceptibility test was carried out and explained according to CLSI 2018 by disk diffusion method and E test method. [8] ATCC 25922, 25923, 27853, 49247,49619, 90028, 35218, 700603, 29213 were used for quality control of indoor antimicrobial sensitivity tests, which were performed weekly.

Statistical analysis
All patient and strain information were stored in the WHONET software. WHONET 5.6 software was used to analyze antimicrobial susceptibility data. pneumoniae (13%). The detection rates of H. influenzae, S. pneumoniae and Moraxella catarrhalis ranked in the top ten, with 8%, 4% and 3% respectively. In fungi, the detection rate of Aspergillus fumigatus was the tenth, accounting for 2%.

Results
A list of other common pathogens was given in fig1 C. The culture results of BALF showed that K. pneumoniae (21%), P. aeruginosa (21%), A. baumannii (17%) and S. aureus (13%) were the most common pathogens. H. influenzae and S. pneumoniae ranked the top ten, with detection rates of 9% and 3%. Among fungi, A. fumigatus and A. flavus ranked the top ten, with detection rates of 4% and 3% respectively. discussion How to evaluate the significance of sputum culture had always been a puzzling 8 problem. In this study, we evaluated the significance of sputum culture from a new perspective. This study compared the difference of pathogen spectrum and antimicrobial sensitivity between sputum samples and BALF, and compared the difference before and after quality control of sputum samples. Studies had shown that, even for qualified sputum specimens, the results of culture and antimicrobial sensitivity were still quite different from those of BALF. BALF was obtained through fiberoptic bronchoscope and could represent lower respiratory tract infection.
However, sputum specimens were easily contaminated by colonies in the upper respiratory tract. Therefore, clinicians needed to be very careful in diagnosing and treating lower respiratory tract infections based on the results of sputum culture.
Sputum specimens were not a good type of specimens, from the point of view of diagnosis of LRIs. Doctors obtained biopsy specimens through fiberoptic bronchoscopy, which could represent LRIs. But after all, it belonged to the invasive operation and was not suitable for every patient. At present, qualified sputum specimens, together with some invasive surgical specimens (transtracheal aspiration, bronchoalveolar lavage, protected brush samples, etc) were acceptable in the global LRIs surveillance project. [3] In the face of LRIs, what kind of specimen to send was another difficult problem. The American Association of Pediatric Infectious Diseases told us that blood cultures should be sent for moderate to severe community-acquired pneumonia in children, especially complex pneumonia.
[9] But at present, the rate of blood culture in Chinese patients with LRIs was not high. A multicenter study from China showed that blood culture isolates accounted for only 5.3% of all specimen types. [1-2] For LRIs, should we do blood culture or sputum culture? The Lancet, an authoritative medical journal, gave us the answer. Different strategies should be adopted for different types of patients. For outpatient 9 blood culture and sputum culture was not needed routinely. For inpatients with low severity, only sputum culture was needed. For inpatient with moderate severity and no ICU sputum culture, blood culture, legionella urinary antigen and pneumococcal urinary antigen should be adopted routinely. For inpatient in ICU with high severity, invasive sampling should also be performed in addition to all the above tests. There were several limitations in the studies. First, in this study, no distinction was made between natural expectoration, induced sputum and sputum aspiration.
Second, whether BALF was a qualified sample had not been judged in this study. We hoped that in future studies, BALF would be interpreted as qualified as sputum specimens. The study protocol was approved by the Tongji Hospital ethics committee for research in health. The Tongji Hospital ethics committee also approved the waiver of informed consent to participate in this study due to its retrospective design. All patient data were anonymous prior to the analysis.

Consent to publish
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.   . Sensitivity rate(%) of Klebsiella pneumoniae from sputum specimens and bronchoalveolar la 15 Figure 5 Susceptibility rate (%) of Staphylococcus aureus to commonly used antibiotics from sputum s