Bacterial etiology of lower respiratory tract infection:a surveillance report from a Chinese large teaching hospital from 2015 to 2019

Sputum is the most common specimen type of lower respiratory tract in China, its cultivation result is easily confused by the bacteria colonized in the oral cavity and pharynx. It is very dicult to evaluate the clinical signicance of sputum culture results both for clinicians and microbiologists. Bronchoscope alveolus lavage uid(BALF)is a good specimen, which can accurately reect the situation of lower respiratory tract infections (LRIs).


Results
The positive rate of BALF culture in our hospital was 18.3% (3467/18935) in 2015-2019. The most common pathogens were Klebsiella pneumoniae (18.1%, 627/3467), Pseudomonas aeruginosa (16.9%, 587/3467) and Acinetobacter baumannii (14.0%, 485/3467). For the eight most common pathogens, 40-70 years old was the highest age of distribution, but for Escherichia coli and Streptococcus pneumoniae, 0-5 years old was also the higher age of distribution. The antibiotic resistance rate of K. pneumoniae to imipenem and meropenem was 30.6% and 30.8%, respectively. The sensitivity of P. aeruginosa to antibiotics other than minocycline and ticarcillin clavulanic acid was all more than 60%. However, the resistance rate of A. baumannii to antibiotics other than tegacyclin and minocycline was all more than 80%.
Conclusions 40-70 years old was the high incidence age of lower respiratory tract bacterial infection. K. pneumoniae resistant to carbapenems (CR-K. pneumoniae) and A. baumannii were a great challenge to clinical treatment and bacterial resistance control.

Background
Lower respiratory tract infections (LRIs) were a leading cause of illness and death in people of all age according to a systematic analysis for the Global Burden of Disease Study 2015 in 195 countries. [1] Bacteria were the main pathogenic factors of LRIs, and the SENTRY surveillance project had carried out etiology and antimicrobial resistance surveillance of LRIs in North America, Europe, Asia Paci c and Latin America. [2] The surveillance results of SENTRY also showed that the etiology of LRIs was different in different regions all over the world. [2] Therefore, the analysis of antimicrobial resistance monitoring data accumulated over many years is of great practical signi cance for the empirical treatment in this region.
Sputum was the most common specimen type of lower respiratory tract in China. Data from China Antimicrobial Resistance Surveillance System (CARSS) in 2015 indicated sputum accounted for 81.6% of the major specimens from inpatients of respiratory departments in 91 general hospitals in seven regions of China. [3] However, the evaluation of LRIs by sputum smear or sputum culture has been controversial.
Results of sputum culture are easy to be confused by the colonized bacteria in the oral cavity and throat, so it is di cult to judge whether the culture result is infection or colonization, so the evaluation of sputum culture has always been a di cult problem for clinicians and microbiologists. As early as the 1970s, microbiologists developed many standards to evaluate the quality of sputum samples. [4] Only quali ed sputum samples have the value of culture. Even for quali ed sputum samples, the results of culture were in uenced by the use of antibiotics in the evaluation of LRIs. [5] Therefore, sputum specimen is not a good specimen type to evaluate LRIs. Bronchoscope alveolus lavage uid(BALF)is a good specimen, which can accurately re ect the situation of LRIs. [6] In order to accurately analyze the pathogenic bacteria and antimicrobial resistance of LRIs in this region, the BALF data accumulated from 2015 to 2019 were analyzed retrospectively, so as to guide clinicians to reasonably select antibiotics for empirical treatment.

Study design and data collection
All analysis samples and strain information were from Tongji Hospital. The culture results of BALF from January 2015 to October 2019 were analyzed retrospectively. The differences of age distribution, seasonal distribution and antimicrobial resistance of main pathogens were analyzed.

Strain identi cation and antimicrobial sensitivity testing
The culture of strains was carried out by conventional methods in strict accordance with the standardized operation procedures of the department. The identi cation of strains was carried out by biochemical experiments, automatic identi cation system (Vitek-2-compact, BioMerier Products) and/or IVD-MALDI Biotyper (Bruker, Germany). Antimicrobial susceptibility test was carried out according to CLSI by disk diffusion method and E test method.

Statistical analysis
All strain information and patient information were stored in WHONET 5.6 software. The results of antimicrobial sensitivity, age distribution and seasonal distribution of the strains were analyzed by whonet5.6 software. The analysis of antimicrobial sensitivity data was based on CLSI 2019 standard. [7] For the same strain from the same patient, only the rst isolated strain was analyzed according to CLSI M39 in order to avoid the in uence of repeatedly isolated strains on antimicrobial resistance statistics. [8] For the antimicrobial sensitivity results of Enterobacteriaceae to carbapenems, when result of disk diffusion method was not sensitive, the method of E test was used to con rm the results and the results of antimicrobial sensitivity test were analyzed according to the results of E test.

Seasonal distribution of LRIs pathogens
Different strains had different epidemic seasons. S. aureus isolated most in spring (February to April), while A. baumannii, H. in uenzae and S. pneumoniae isolated most in summer (May to July). For K. pneumoniae, P. aeruginosa and S. maltophilia, autumn (August to October) was the season with the most isolates, while E. coli was in winter (November to January) with the most isolates. ( g 4) Antimicrobial susceptibility testing For K. pneumoniae isolates, sensitivity rates to cefotaxime, ceftazidime and cefepime were 56.5%, 52.6% and 54.9%, respectively. But resistance rates to imipenem and meropenem were 30.6% and 30.8%, respectively. ( g 5a) The sensitivity of P. aeruginosa to minocycline and ticarcillin clavulanic acid was 35% and 40.6% respectively, and the sensitivity to other commonly used antibiotics was all more than 60%. ( g 5b) The sensitivity rates of A. baumannii to tegacyclin and minocycline were 42.2% and 32.9% respectively, and the resistance rates to other antibiotics were all higher than 80%. ( g 5c) The sensitivity rates of MRSA to vancomycin, teicoplanin and linezolid were all 100%, to tegacyclin, rifampicin and Trimethoprim Sulfamethoxazole were 92.7%, 72.5% and 96.7%, respectively, but resistance rates to other antimicrobial agents were more than 50%. ( g 5d) The resistant rate of MSSA to penicillin was 90.9% and the sensitive rates to erythromycin and clindamycin were 56.9% and 78.4%, respectively. The sensitivity rates to other antibiotics were more than 85%. ( g 5e) The sensitivity rate of H. in uenzae to ampicillin was 57.9%. The resistance rate to Trimethoprim Sulfamethoxazole was 54.5% and the non-sensitivity rates to azithromycin, cipro oxacin and cefotaxime were 18.2%, 3.7% and 4%, respectively. ( g 5f) The sensitivity rates of S. maltophilia to Trimethoprim Sulfamethoxazole, minocycline and levo oxacin was more than 80%. ( g 5g) For E. coli isolates, sensitivity rates to cefotaxime, ceftazidime and cefepime were 50.4%, 24.4% and 26.0%, respectively. But resistance rates to imipenem and meropenem were both 2.4%. ( g 5h) The sensitivity of S. pneumoniae to penicillin was 41.8% and 94.5% according to the breaking points of meningitis and non-meningitis, respectively. And the sensitivity rate to ceftriaxone was 66.4% and 89.1% according to the breaking points of meningitis and non-meningitis, respectively. The nonsensitivity rate of S. pneumoniae to oxacillin was 56.4%. ( g 5i) Trends of multi-drug resistance (MDR) strains The detection rate of MRSA was decreasing year by year. The detection rate of A. baumannii resistant to carbapenems was always at a high level, all of which were more than 80%. Different from A. baumannii, the detection rates of carbapenem resistant P. aeruginosa has been lower than 25% from 2015 to 2019. The detection rate of carbapenem resistant E. coli was lower than 6%, but the detection rate of carbapenem resistant K. pneumoniae was 15% -40%. ( g 6)

Discussion
Our surveillance data showed that K. pneumoniae, P. aeruginosa and A. baumannii were the most common pathogens of LRIs. Our data was different from the results of a global multi center report from the SENTRY Antimicrobial Surveillance Program (1997-2016), which showed in Latin America, the most common pathogens were P. aeruginosa, S. aureus and Acinetobacter, while in Europe, Asia Paci c and North America, the most common pathogens were S. aureus, P. aeruginosa and K. pneumoniae. [2] S. aureus ranked fourth in our area. Our pathogenic spectrum was consistent with the data reported by CHINET (China antimicrobial surveillance network) of China in 2018. [9] In the developed world, the United States a study indicated S. pneumoniae, S. aureus and Legionella pneumophila were the most common pathogens of community-acquired pneumonia in adults (≥ 18 years old) from 2010-2012. [10] However, the most common pathogens of adult community-acquired pneumonia in Malawi, a developing country in Africa were S. pneumoniae, Mycobacterium tuberculosis and non-Mycobacterium tuberculosis. [11] We found that in different regions of the world, the pathogenic bacteria of LRIs were signi cantly different. Therefore, it was of great signi cance to analyze the distribution of local pathogens for epidemiological research and clinical experience.
This study found that the main population of LRIs was 40-70 years old, for the eight most common pathogens K. pneumoniae, P. aeruginosa, A. baumannii, S. aureus, H. in uenzae, S. maltophilia, E. coli and S. pneumoniae. For E. coli and S. pneumoniae, the age range of 0-5 years was also a major distribution. According to data from 195 countries around the world, S. pneumoniae and H. in uenzae type b were the main pathogens of LRIs in children under 5 years old. [1] With the global promotion of S. pneumoniae and H. in uenzae b vaccine, their incidence had decreased signi cantly. [12] Among all infectious diseases, vaccines were the most effective means of control. Our study found that the age distribution of pathogens in LRIs was different from that in bloodstream infection. Studies from Malawi, Africa, have shown that K. pneumoniae, which caused bloodstream infection, mainly came from the ages of 0-5 and 75-79. [13] To understand the age distribution of the main pathogens was helpful for clinicians to choose the appropriate antibiotics when they took experiential treatment. For example, in the treatment of children, uoroquinolones should be avoided because of their in uence on bone development. For the elderly patients with poor liver and kidney function, the liver and kidney toxicity of antibiotics should be considered in the empirical treatment.
A systematic analysis on global patterns in monthly activity of in uenza virus, respiratory syncytial virus, parain uenza virus, and metapneumovirus found in uenza virus had clear seasonal epidemics in winter months in most temperate sites but timing of epidemics was more variable and less seasonal with decreasing distance from the equator and other viruses had obvious epidemic seasons. [14] In our data, S. aureus had the highest isolation rate in spring, A. baumannii, H. in uenzae and S. pneumoniae had the highest isolation rate in summer, K. pneumoniae, P. aeruginosa and S. maltophilia had the highest isolation rate in autumn, while E. coli had the highest isolation rate in winter. It is of great signi cance to understand the seasonal prevalence of pathogens in LRIs for disease prevention and control as well as vaccination. [15] This study showed that K. pneumoniae was the main pathogen of LRIs. Cephalosporin was often used in the treatment of LRIs because of its small side effects. This study showed that the sensitivity rates of K. pneumoniae to ceftazidime and cefotaxime were 56.5% and 52.6% respectively, and the sensitivity rate to cefepime was 54.9%. The main resistance mechanism of K. pneumoniae to cephalosporin was the expression of extended spectrum β -lactamase. (ESBLs). [16] When cephalosporins were resistant, carbapenems were often considered in clinical experience. Our data showed that the resistance rates of K. pneumoniae to imipenem and meropenem were 30.6% and 30.8%, respectively. Large data analysis from China's multi centers shows that the antimicrobial resistance rate of K. pneumoniae to imipenem and meropenem increased year by year from 2005 to 2018, and the resistance rates of meropenem and imipenem increased from 2.9-26.3% and 3-25%, respectively. [9] However, the situation in Germany was different from that in China. Monitoring data from the German Antimicrobial Resistance Surveillance (ARS) show that the isolation rate of K. pneumoniae, which was not sensitive to carbapenems, in Germany in 2011-2016 was only 0.63%. [17] In addition, K. pneumoniae, which was not sensitive to carbapenems, was often resistant to other antibiotics, such as gentamicin, Sulfanilamide and tegacyclin.
[17] Some studies had shown that carbapenem-resistant hypervirulent K. pneumoniae may have a clonal distribution in the hospital. [18] The increase of resistance rate of carbapenems may be related to the unreasonable use of carbapenems. [19] Therefore, the control of carbapenem resistant K. pneumoniae in our hospital was the top priority for infection control.
This study found that in addition to K. pneumoniae, another strain with high resistance rate was A. baumannii. In addition to tegacyclin and minocycline, the resistant rates to other commonly used antibiotics of A. baumannii were more than 80%. Data from CHINET in 2018 shows that A. baumannii isolates from blood, cerebrospinal uid and lower respiratory tract had a very high resistance rate to commonly used antibiotics (almost all more than 50%, even more than 80%). [9] At present, the antimicrobial resistance of A. baumannii in China had reached a very serious level. The high resistance of A. baumannii may be related to the expression of carbapenemase ox-23, ox-24 and ox-51. [20] The formation of bio lm was bene cial to the long-term existence of A. baumannii in the environment and it was found that there was homology between the infection strains of patients and the colonization strains in the environment, indicating that the highly resistant strains were cross transmitted between patients and the environment. [21] Therefore, the control of A. baumannii must rely on the strengthening of hospital infection control measures.
There are several limitations in this study. First of all, the types of LRIs such as community acquired, hospital acquired and ventilator related infection types were not distinguished in this study. Secondly, the molecular epidemiology of carbapenem resistant K. pneumoniae and A. baumannii were not analyzed.

Conclusions
At present, the primary task of LRIs is to control the spread of K. pneumoniae which are resistant to carbapenem and A. baumannii. 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.

Competing interests
The authors declare that they have no competing interest.

Author Contributions
LT analyzed the data and ZZ wrote the article.  The main pathogens isolated from lower respiratory tract infection strati ed by age  The distribution proportion in the four seasons of the main pathogens isolated from LRIs