Pneumonia causes high mortality among children under 5 years old. According to data from the World Health Organization, there are more children died from pneumonia than AIDS, malaria and measles combined (Adegbola, 2012). In industrialized countries, 1–10% of community-acquired and 20–50% of nosocomial bacterial pneumonia in children were caused by S. aureus (Bradley, 2005; Chisti et al., 2009). First reported in 1961, soon after the clinical introduction of methicillin, MRSA (methicillin-resistant Staphylococcus aureus) has become major nosocomial pathogen worldwide, and was found in the community (community-associated MRSA, CA-MRSA) in the late 1980s. Four pediatric necrotizing pneumonia caused by CA-MRSA in 1999 raised a global public health concern (Centers for Disease Control and Prevention, 1999). Now it has been estimated that up to 40% of patients with CA-MRSA had severe or fatal pneumonia (Wallin et al., 2008; Carcillo et al., 2009).

Several CA-MRSA clones have been reported with different regional prevalence, while ST30 clones are found worldwide (Aires De Sousa et al., 2003; Hsu et al., 2006). ST8 (USA300) and ST1 (USA400) are mostly reported in the United States and Canada (Mulvey et al., 2005; Tenover et al., 2006), ST80 is the prevalent clone in Europe (Holmes et al., 2005; Wannet et al., 2005; Fossum and Bukholm, 2006), and ST59 spreads mostly in the Asian-Pacific area (Coombs et al., 2004; Ho et al., 2007; Huang and Chen, 2011).

In recent years, some researchers reported MRSA infections in Chinese children. ST239 and ST59 were common clones in hospital- associated infections caused by MRSA (Ning et al., 2015). However, rare data of prevalence, antimicrobial susceptibility and molecular epidemiology of MRSA causing childhood pneumonia were available in Shanghai. We conducted a study on the resistance phenotype and sequence types (STs) of S. aureus isolates from two pediatric hospitals in Shanghai.

The morbidity of childhood pneumonia caused by S. aureus particularly CA-MRSA has been increasing over the past two decades (Gonzalez et al., 2005; David and Daum, 2010). Among the 107 pediatric S. aureus pneumonia patients enrolled in current study, 49 isolates were CA-MRSA according to the definition by J. A. Otter et al and the epidemiological data in Asian pacific area (Ho et al., 2012; Otter et al., 2013; Xiao et al., 2013; Zhang H. et al., 2015). Unlike similar study conducted in Shanghai in 2005, which showed ST239, ST910, and ST88 were the main endemic S. aureus clones causing infections (Zhang et al., 2009), we found that CC59 was the most common CC, and it was also the most common CC among MRSA isolates in Shanghai. In 2009, Chen et al found HA-MRSA stains were still the major pathogen in healthcare- associated bloodstream or secondary to thermal injury in audit in Shanghai, and two main MRSA clones (ST239 and ST5) were prevailing in patients (Chen et al., 2012, 2013). ST239 was also reported as an epidemic clone causing S. aureus surgical site infections in orthopedic patients in Shanghai in 2011 (Gu et al., 2015a). Nevertheless, ST239 was not detected in our study, and CC8 (ST8, ST630, ST1821) was discovered in only 3 isolates. It seemed that ST59 had replaced ST239 in terms of dissemination in Shanghai. STs which proved capable of developing high-level of resistance to fluoroquinolones without suffering significant fitness cost would supplant others. However, other factors would determine clonal shifts in facilities/wards where fluoroquinolones are inappropriate in use (Fuzi, 2016). The dominant MRSA sequence type replacements are worth further studies. A study performed in 2008 showed ST59-SCCmec IV-t437 was the predominant clone of CA-MRSA which caused community –onset childhood pneumonia in China (Geng et al., 2010); however, the present study suggested that ST59-SCCmec IV-t172 had replaced ST59-SCCmec IV-t437 as the most common clone in CA-MRSA in Shanghai. Gu's study implied CC59 was the most common CC among adult patients with skin and soft tissue infections (SSTIs) in Shanghai in 2011 (Gu et al., 2015b). The study conducted among nursing home residents in Shanghai in 2014 showed CC1 was the most common clonal of S. aureus carriage (Zhang J. et al., 2015). This might indicate that CC59 is common in SSTIs and pneumonia, and will not be affected by population factors. CC398 was also recently reported as a livestock-associated clone in skin and soft-tissue infections (SSTIs) in China (Gu et al., 2015b) and the spa type t034 (CC398) is a typical livestock-associated spa type. In Europe, patients carrying this spa type are usually in contact with major animal reservoir (mostly pigs) carrying these MRSA (Köck et al., 2009). Livestock-associated S. aureus CC398 (ST398 with t034) was found in 7 isolates (4MRSA and 3 MSSA), and the SCCmec type of all CA-MRSA was V in this study. Clinical data on any possible children contact are not available in this study, but all children infected by ST398-t034 S. aureus were under 3 months. Transmission between animals and humans requires further research. ST188 (14/58, 24.1%) was the common genotype of MSSA in S. aureus pneumonia in Chinese children, which was in agreement with the results reported by Qiao et al. (2014).

The prevalence of enterotoxins among S. aureus (57.9%, 62/107) isolated from pediatric patients, is actually consistent with that from adult (53.8–65.9%) in the same region (Chen et al., 2013; Gu et al., 2015a,b). ETs positive strains in MRSA and MSSA were 35 (71.4%, 35/49) and 27 (46.6%, 27/58) respectively. The prevalence of ETs in MRSA was lower than the occurrence of MRSA from surgical site infections (80.1%, 29/36) and SSTI (92.9%, 13/14) patients and higher than bloodstream infection (59.7%, 37/62) patients as our previously studies in Shanghai (Chen et al., 2013; Gu et al., 2015a,b). MRSA tended to carry the seb and sea genes, while MSSA tended to carry seg and sei genes, which was similar with the findings of our study on S. aureus infecting surgical site in orthopedic patients (Gu et al., 2015a). Further studies are warranted to get better understanding of this epidemic phenomenon in Shanghai.

Panton-Valentine leukocidin (PVL) is a bicomponent toxin, causing the lysis of leucocytes by forming a pore in their membrane, and it is a main virulence factor of S. aureus, independent of methicillin resistance (Gillet et al., 2011), and responsible for severe necrotising pneumonia. In 2006, Geng et al. found 22 MRSA (22/50, 40%) stains that were PVL- positive from children with community-onset pneumonia in China. In the same year, Han et al isolated 9 MRSA strains harboring lukS/F-PV from 8 infants with pneumonia and 1 adult with prostatitis in Shanghai (Han et al., 2010). In Qiao's study the pvl gene was detected in 27% of the isolates in children with invasive CA-SA infections from 2011 to 2013 in China (Qiao et al., 2014). In current study, a total of six lukS/F-PV -positive isolates were identified, and it was lower than the occurrences of lukS/F-PV-positive S. aureus previously reported (Geng et al., 2010; Han et al., 2010; Qiao et al., 2014). Three of lukS/F-PV -positive isolates in CA-MRSA were ST59 and the remaining three were ST6, ST217, and ST398 MSSA respectively.

When CA-MRSA infections is suspected the initial empiric antibiotic therapy including vancomycin or Clindamycin will be applied according to the American Academy of Pediatrics' Committee on Infectious Diseases' recommendations in 2009 (American Academy of Pediatrics Committee on Infectious Diseases: Staphylococcal Infections et al., 2009). Current study showed linezolid and vancomycin were more susceptible. The results of antimicrobial susceptibility test suggested that linezolid and vancomycin were appropriate antibiotics for treating S. aureus including CA-MRSA childhood pneumonia. CA-MRSA strains are resistance to β-lactams and cephalosporins and mostly susceptible to several non-β-lactam antibiotics (Shilo and Quach, 2011; Otter and French, 2012). Surprisingly MRSA in our study revealed low susceptibility to erythromycin (18.4%) and clindamycin (40.8%), some of which are inducible resistance. So Clindamycin may not be the most suitable empirical treatment of CA-MRSA childhood pneumonia in Shanghai. All Seven isolates susceptible to penicillin were β-lactamase producers, precluding penicillin in their treatment. Fluoroquinolones are unsuitable drug for children with pneumonia in the pediatric wards. This may be an explanation of the relatively low rate of resistance to ciprofloxacin in MRSA isolates. What is noteworthy is that all ST59-SCCmec IV-t172 were ciprofloxacin-intermediate. The microbiology laboratory study is warranted for development of an effective antimicrobial regimen in this area.

Conceived and designed the experiments: LH and PH. Performed the experiments: ZS and FG. Analyzed the data: YN. Contributed reagents/materials/analysis tools: LH, PH, and XG. Wrote the paper: ZS, FG, and LH.

This study was supported by National Natural Science Foundation of China (grant numbers 81472010 and 81471908). This study was also supported by the Shanghai 3-Year Plan of the Key Subjects Construction in Public Health-Infectious Diseases and Pathogenic Microorganism (grant number 15GWZK0102).