Hepatocyte growth factor (HGF) was originally discovered as a mitogen of hepatocytes in adult rats and exhibits a cytotoxic effect in certain tumor cells, such as those in hematoma (1-2). HGF is a paracrine factor that is secreted by mesenchymal cells, such as fibroblasts, bronchial epithelial cells, alveolar macrophages and neutrophils, and plays an important role in lung development, inflammation and tissue repair (3-5). The biological effect of HGF is mediated by the cell membrane receptor MET (6). HGF synthesis is regulated by prostaglandins, cytokines and hormones, while its cellular secretion is stimulated by acute phase proteins, such as tumor necrosis factor-α (TNF-α), interleukin (IL)-1α, IL-1β, IL-6, platelet derived growth factor and fibroblast growth factor β, as well as cyclic adenosine monophosphate (7-10). HGF can be detected in edema fluid during acute lung injury, where it is produced locally by fibroblasts and inflammatory cells (11-14). HGF is therefore used to assess progression of Legionella pneumonia (15), solitary pulmonary nodules (16) and bacterial pneumonia (17).

Mycoplasma pneumoniae pneumonia (MP), which is unresponsive to commonly used antibiotics such as penicillin, is caused by a microorganism associated with atypical pneumonia (18). Mycoplasma pneumoniae infections vary dramatically in clinical presentation, ranging from mild and self-limiting upper respiratory symptoms to radiographically confirmed pneumonia that requires hospitalization (19). In some cases, Mycoplasma pneumoniae infection may result in severe clinical consequences that involve organs other than lungs, such as encephalitis and acute hepatitis, which is known as extrapulmonary complications (20, 21). As the Mycoplasma pneumoniae incubation period can last as long as 3 weeks, and prolonged shedding follows infection, outbreaks often go unnoticed and the course of infection can extend for long periods of time (22). The clinical detection of Mycoplasma pneumoniae relies on serological IgM testing or PCR assay; culture is rarely used because the microorganism grows very slowly (23, 24). Macrolides are the treatment of choice for MP, while tetracyclines and fluoroquinolones also prove to be effective (25).

Pneumonia treatment is particularly important in children, as it remains the most common cause of mortality and morbidity in children <5 years of age globally (26). Although preventable, pneumonia can lead to chronic lung conditions, such as pulmonary fibrosis, if it is recurrent or untreated (27). A study among a Finnish population reported that MP was detected in 30% of pediatric community-acquired pneumonia patients and in >50% children aged 5 years or older (28).

In the present study, serum HGF levels were assessed in MP as well as in bacterial pneumonia.

There were no significant differences between age and sex of the subjects in the control, bacterial pneumonia and MP groups. Additionally, there were no significant differences between ALT and AST levels among the three groups (Table I). CRP levels significantly differed among bacterial pneumonia, MP pneumonia and control subjects (P<0.001; Table I). As indicated in Fig. 1A, serum HGF levels were significantly higher in the MP and bacterial pneumonia groups compared to levels in control subjects on the day of admission before treatment (P<0.05). HGF levels did not vary significantly between MP and bacterial pneumonia (Fig. 1A). Similarly, CRP levels did not vary significantly between MP and bacterial pneumonia (Fig. 1B), while LDH levels are significantly higher in MP pneumonia compared with bacterial pneumonia (P<0.05; Fig. 1C). Patients who suffered from MP were divided into severe MP (n=27) and non-severe MP (n=38), according to the guidelines of community-acquired pneumonia management in children in China (30). HGF was significantly higher in severe MP than in non-severe MP on the day of admission (P<0.05; Fig. 2A). In addition, CRP was significantly higher in severe MP [17.8 (5.5-30.1) mg/l] than in non-severe MP [5.9 (2.4-25.0) mg/l; Table II; Fig. 2B] as was LDH (337.0±120.3 vs. 256.9±54.4 IU/l; Table II; Fig. 2C). There were no obvious differences in sex, WBC count and liver function (as indicated by ALT and AST) between severe cases and non-severe cases. However, the average age of severe MP patients was significantly older than that of non-severe patients (Table II).

In those 42 patients who responded to 2 days of azithromycin treatment, serum HGF levels dropped significantly (P<0.05; Fig. 3A) as did CRP levels (P<0.05; Fig. 3B). In addition, LDH was also significantly reduced (P<0.05; Fig. 3C). There was no evident change in HGF, CRP or LDH (P>0.05) in the 23 patients who were not responsive to azithromycin treatment (Fig. 3).

Finally, the sensitivity and specificity of CRP, LDH and HGF in assessing effectiveness of azithromycin treatment were compared. As indicated in Fig. 4, HGF showed advantage over CRP and LDH in evaluating azithromycin treatment with higher sensitivity or specificity (P<0.05). The sensitivity (97.6%) and specificity (78.3%) levels of HGF were both higher than those of CRP (91.2% sensitivity; 69.1% for specificity) and LDH (85.1% sensitivity; 70.2% specificity), respectively (Table III).

HGF is known to be elevated in Legionella pneumonia (15) and bacterial pneumonia (17), which led to the hypothesis that it may be elevated in MP. When children are infected with Mycoplasma pneumoniae IL-2, IL-6 and TNF-α are released into the peripheral blood (32), which can stimulate secretion of HGF (10). HGF is also known to be secreted at high levels in viral pneumonia and HGF/MET signaling enhances the innate immune response, by stimulating the secretion of IL-8 and granulocyte macrophage colony-stimulating factor (33). Additionally, both HGF and CRP levels correlated with the severity of MP, as shown in the present study. HGF has been used to assess the severity of avian origin influenza A (H7N9) virus and worked as an independent outcome predictor (34). Therefore, HGF has been suggested to act as a general indicator of activity of the innate immune response, rather than as a specific marker for MP. However, this does not impede its application in assessing MP. Interestingly, the present study showed that LDH was also upregulated in severe MP. In the present study LDH was used as a biomarker to predict refractory MP, which is characterized by prolonged fever and hospital stay as well as more severe pulmonary symptoms in children (35) as an alternative to erythrocyte sedimentation rate and CRP (36). LDH may be an indicator of severity of MP. In the present study, the mean age of severe MP patients was higher than non-severe ones, which may be due to the more mature immune system in older children causing longer duration of fever, higher CRP, and more severe pulmonary lesions compared to younger patients.

Determining the severity of pneumonia is not always straightforward. The World Health Organization criteria divide the severity of pneumonia into 2 grades: Severe pneumonia and very severe pneumonia (37). Although there is an association between these classifications and clinical outcome (38,39), it is still a relatively coarse classification system and does not consider population variation due to geography. As all subjects in the present study were Chinese children, Chinese national guidance was applied to divide pediatric pneumonia into severe and non-severe cases (30). Identifying severe cases of MP is important, because the use of steroids is advocated alongside antibiotics for such cases (30). Macrolides have been shown to be effective in severe cases of Mycoplasma pneumoniae infection (25). The results of the present study suggest that HGF levels could be assessed to determine whether children with MP require early steroid treatment.

β-lactam antibiotics, including penicillin and cephalosporins, are considered to be the drugs of choice for treating respiratory diseases, as they are active against most respiratory bacterial pathogens (40). However, they are ineffective against Mycoplasma pneumoniae because they target the cell wall, while Mycoplasma pneumoniae lacks a cell wall (41). Macrolides and tetracyclines, which act as protein synthesis inhibitors, and fluoroquinolones, which act against topoisomerases to inhibit DNA synthesis and replication, are usually highly effective against Mycoplasma pneumoniae and are the drugs of choice (25). However, only macrolides are recommended for children, as tetracycline and fluoroquinolone administration can lead to a number of age-related adverse events, such as cardiotoxicity and inhibition of bone growth (42,43). Thus, in the present study azithromycin, a macrolide, was used to treat pediatric patients. However, approximately one third (23/65) of patients did not respond to this treatment. Since the first report of macrolide resistance in 2001, it has increased rapidly, with the highest prevalence in East Asia (44,45). The mechanism underlying macrolide resistance is related to point mutations in domain V of the 23S ribosomal RNA gene of Mycoplasma pneumoniae and can usually be detected with the onset of the disease (46). CRP is the only factor demonstrated to significantly predict the efficacy of macrolides in MP patients (47). The results of the present study suggest that HGF may also be useful to assess the efficacy of azithromycin administration to MP patients. HGF has previously been used to evaluate the effectiveness of treatment in patients with acute brucellosis and acute myocardial infarction (48,49). Though this suggests that HGF is a general indicator for inflammation or the immune response the results of the present study indicated that HRP had a higher sensitivity and specificity than CRP in assessing efficacy of azithromycin treatment of MP.