The distribution of pathogens associated with CAP is attracting world-wide attention. The present study is the first in China to prospectively investigate the pathogen distribution in a large, broadly distributed patient population. The study comprised several representative districts and covered all seasons of the year. The data reveal a unique pattern of pathogen distribution in CAP patients in China and show a relatively higher susceptibility of S. pneumoniae isolates to penicillin but relatively lower susceptibility to fluoroquinolones than that reported in Europe and North America.
In previous studies, 30 to 60% of the CAP cases did not yield an identifiable pathogen , in line with our results. Although S. pneumoniae remains the most prevalent or frequently isolated etiologic agent in cases of CAP [2, 13], other organisms, such as H. influenzae and M. catarrhalis, as well as the so-called atypical pathogens, including M. pneumoniae, C. pneumoniae, and L. pneumophila, are now being reported more frequently than in the past [14, 15]. Atypical pathogens are responsible for 30 to 40% of cases of CAP . In some reports, the infection with atypical pathogens was more than with S. pneumoniae, especially for M. pneumoniae. Moola et al  reported that when pathogens were confirmed in 131 of 504 patients, M. pneumoniae was found in 25% patients, and S. pneumoniae in 22%. In a study performed in Spain , M. pneumoniae was the most frequent pathogen (33%), with S. pneumoniae being found in only 19 of 110 patients with CAP.
In our study, the rate of atypical pathogen infection is 31.3% (if convalescent sera obtained longer than 4 weeks after the acute specimen, L. pneumophila may be more, and this percentage may be higher); M. pneumoniae is the most prevalent etiologic agent, accounting for 20.7% of CAP cases, i.e. more than those associated with S. pneumoniae. Such finding is similar to our previous study between 2001 and 2002 in Beijing [9, 17]. Previous reports show that viral infection in patients with CAP varies from 4% to 39% , the most frequent being influenza spp. In the present study, respiratory viruses account for 19.0% of tested patients, with adenovirus being the most common.
Previous studies have shown that some patients with CAP can have mixed infections involving both bacterial and atypical pathogens [2, 19]. C. pneumoniae is frequently found in mixed infections and has been detected in over half of cases with S. pneumoniae in reports from other countries [2, 19]. In our study, however, M. pneumoniae is the most frequent co-infecting pathogen. Multiple simultaneous infections might interfere with the pulmonary cleansing function and thus help establish setting for CAP . Pathogens such as C. pneumoniae induce ciliostasis in human bronchial epithelial cells , and M. pneumoniae exerts a toxic effect on ciliated human epithelium . Currently, the clinical implications of mixed infections are still undetermined. Influenza and other viruses can cause primary viral pneumonia; secondary bacterial infections are common in hospitalized adults, and the reported frequency has ranged widely, from 26% to 77% in different studies . The common cause of bacterial superinfection is S. pneumoniae. In our study, mixed infection of respiratory viruses with other pathogens accounts for 12.5%. The most common co-infecting pathogen was M. pneumoniae.
The impact of age is restricted to an association of younger patients (age ≤ 50 years) with M. pneumoniae. This finding is consistent with several previous studies [23, 24], which show that M. pneumoniae is the most common etiologic agent in the 17–44 year age group. Patients without co-morbid illnesses are also more likely to have M. pneumoniae infection. In our study, only 27.8% patients of M. pneumoniae have co-morbid illnesses. It has been reported that C. pneumoniae infection is more likely to occur in elderly persons with co-morbid disease than in those who are otherwise healthy . However, our data showed a higher incidence of C. pneumoniae infection in older patients without co-morbid illness than with, contradicting previous conclusion. An association of older age with Gram-negative pathogens has been observed by several groups [12, 26]. It is well known that colonization of the oropharyngeal mucosa by aerobic Gram-negative bacilli increases with age . In our study an age of > 50 years was significantly associated with these pathogens. Pulmonary co-morbidity also predisposes to pneumonia due to K. pneumoniae or P. aeruginosa, probably due to previous tracheabronchial colonization in COPD . In our study, Staphylococcus aureus was more frequent in PORT I than in PORTs II through IV, probably due to patients having low co-morbidities of diabetes mellitus and to sample size being too small.
Antibiotic resistance in pneumococci is now considered to be a global problem. The ANSORP (Asian network for surveillance of resistant pathogens) study, conducted in Asian countries from September 1996 to June 1997, showed that the non-suscepibility to penicillin was as high as 79.7% in Korea and non-susceptibility to erythromycin was as high as 90.5% in Taiwan . In China, penicillin-non-susceptibile S. pneumoniae ranged from 13.9% to 42.7%, and erythromycin-non-susceptibile S. pneumoniae ranged from 42.5% to 83.6% in surveys conducted from 1997 to 2003 [5, 6, 14, 30–32]. Among the 63 strains of S. pneumoniae obtained from the present study, non-susceptibility to penicillin was 22.2%; non-susceptibility to azithromycin was 79.4%. In 2008 year, CLSI changed the break point of penicillin for patients with non-CNS infections, according to this, the non-susceptibility to penicillin was only 3.2%. Although new fluoroquinolones have good activity against S. pneumoniae, the resistance rate of new fluoroquinolones is higher in China than that reported in other countries except for Hong Kong , which is possibly a result of selective pressure due to increased quinolone use in China.
American Thoracic Society guidelines  advocate that all populations with CAP should be treated for possible infection with atypical pathogens. Macrolides are recommended as the first-line choice for outpatients. However, such a recommendation may not be suitable for patients in China because the prevalence of macrolide resistance is so high (about 70% among S. pneumoniae isolates) and because ermB-mediated high-level resistance is more frequent than low-level resistance conferred by mefA (79.1% contained the ermB gene, 10.8% contained the mefA gene, and 10.1% harbored both the ermB and mefA genes) as shown in our previous study . In the present study, for patients infected with M. pneumoniae and/or C. pneumoniae, the cure rate with β-lactams plus macrolides or with fluoroquinolones was not significantly better than with β-lactams alone. A meta-analysis also showed no advantage of antibiotics active against atypical pathogens over β-lactams . Thus, β-lactams (penicillin, amoxicillin, cephalosporins) should remain agents of choice in the initial management of mild CAP without co-morbidity in China. But for moderate or severe infection or in patients with co-morbidities, therapy should be recommended with either a β-lactam/macrolide combination or an anti-pneumococcal fluoroquinolone alone [36–39].