Introduction

Chronic rhinosinusitis (CRS) is an inflammatory disease of the nasal and paranasal sinus mucosa, which is considered as one of severe chronic health problems in Western and Asian countries. CRS can be categorized into several subtypes according to various symptoms of sinus inflammation with the time period of at least 12 weeks [1, 2]. The underlying mechanism of the pathogenesis of CRS is complicated and remains unclear. Local and systemic factors, as well as the microbiological, environmental, genetic and iatrogenic factors play important roles in the pathophysiology of CRS. CRS can be considered as a functional abnormality of the host and environment interaction that occurs in contact site of the body and the nasal sinus mucosa [3].

A number of studies have showed that microflora play an important role in promoting the development of early immune function, maintaining immune balance, removing pathogens, and regulating infection and allergic disease sensitivity [4,5,6]. Recent studies of the airway microflora have demonstrated that the nasal sinuses are not sterile in the healthy state [7,8,9,10,11], indicating that the composition of the microbial community may be a potential regulator of the inflammation progression of CRS [12,13,14,15,16,17]. Previous studies have reported that patients with chronic inflammatory airway disease have unique microbial characteristics, which suggest the severity of the disease [18, 19]. Another study showed that the sinus microflora of CRS patients exhibit significantly reduced bacterial diversity compared with that of healthy controls [8]. They demonstrated that the multiple, phylogenetically distinct lactic acid bacteria were depleted while the Corynebacterium tuberculostearicum was significantly increased in the cohort of CRS patients. Similarly, CRS patients with altered microflora composition and greater abundance of Staphylococcus aureus were also reported in the United States [11]. It is worth noting that the bacteria colonizing the nasal airways of Chinese CRS patients and Caucasian CRS patients are not similar. In Western countries, CRS is classified as CRSsNP, a Th1 polarized disorder, or CRSwNP, a Th2 polarized disorder with eosinophilic inflammation as a key feature of White patients with CRSwNP [20]. In contrast, polyps from some Asian CRSwNP patients exhibit Th1, Th17, and KCN cytokine profiles and S. aureus appears to be less common with lower isolation rates than that in Caucasian CRSwNP patients [21]. Additionally, the treatment strategy may be inappropriate in Asian patients with neutrophilic interleukin (IL)-17 biased polyps despite the treatment of eosinophils as the first-line therapy for NPs in the Western countries. Furthermore, the sample size of CRS patients reported in the previous studies is relatively small [8].

Therefore, the aim of this study was to compare the microbiological features in middle meatus samples from Chinese patients with CRS with nasal polyps (CRSwNP), without nasal polyps (CRSsNP), and control subjects, and to further analyze the effect of host-related factors on the distribution of microflora.

Materials and methods

Study population

This was a retrospective study. A total of 251 patients who underwent endoscopic surgery in Beijing TongRen Hospital were enrolled between January 2014 and January 2016. According to EPOS2012 diagnostic criteria [3], 136 cases were diagnosed with CRSwNP (male: n = 89, female: n = 47; mean age: 45 years), and 66 cases were diagnosed with CRSsNP (male: n = 35, female: n = 31; mean age: 42 years). Forty-nine control patients with nasal congestion and snoring (without chronic rhinosinusitis) were selected and they underwent nasal cavity expansion surgery. The mean age of control subjects was 41 years (male: n = 31, female: n = 18). All patients underwent sinus CT examination and some received sinus MRI examination before operation. Other information on allergic rhinitis, asthma, the percentage of eosinophils (EOS) in peripheral blood, and history of smoking and surgery was collected. Antibiotics and glucocorticoids were avoided at least 1 month before surgery. Preoperative diagnosis or suspected fungal ball sinusitis, immotile-cilia syndrome and cystic fibrosis, other immune dysfunction, and pregnant and lactating patients were excluded from the study. This study protocol was approved by local ethics committee of Beijing TongRen Hospital. Patient consent was not required because of the retrospective nature of the study.

Sample collection

The specimens were collected by preoperative nasal endoscopy or endoscopic sinus surgery (ESS). Before the collection, the nasal vestibular area was disinfected with Anne iodine. The sterile swabs (Jinzhang, Tianjin, China) were endoscopically guided to the middle meatus, rotated at least three full turns. If the patients experienced significant swelling in their inferior turbinate, ephedrine was used to contract the inferior turbinate, and avoided other parts of the body from contaminating the swabs. Each specimen was labeled with complete patient information including ward, disease number, name, gender, age, diagnosis, and collection site and time. Subsequently, the swabs were placed in a transfer medium immediately to prevent drying (Jinzhang, Tianjin, China). All samples were transferred to the microbial laboratory within 30 min after the collection.

Culture and identification of microflora

All samples were subjected to normal bacterial culture, anaerobic culture and fungal culture.

Normal bacterial culture: the specimens were inoculated in Colombian blood agar medium and Chlamydomonas Glabriate agar medium (BioMérieux, China), incubated for 24 h at 35 °C under 5% CO2; Specimens were inoculated in the MacConkey Agar Medium (BioMérieux, China) and incubated for 24 h at 35 °C.

Anaerobic culture: the specimens were inoculated in the blister medium (Jinzhang, Tianjing, China) and incubated for 48 h at 35 °C under anaerobic conditions.

Fungal culture: the specimens were inoculated in the Soba dextrose (chloramphenicol) medium (BioMérieux, China), and the fungal growth was monitored at 2, 7 and 14 days after the inoculation. Microbiological isolation and identification were performed according to the methods defined in the Manual of Clinical Microbiology [22].

Statistical analyses

All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC). A P < 0.05 (two-tailed) was considered statistically significance. Comparisons for categorical variables were performed by the Pearson Chi-square test or Fisher’s exact test. Comparisons for continuous variables were performed using ANOVA or the Kruskal–Wallis test, and multiple comparisons was performed using the Student–Newman–Keuls (SNK) method. Comparisons between two groups for continuous variables were performed by an either t test or Mann–Whitney test.

Results

Table 1 presents the characteristics of the three groups. There were no significant differences in sex, age, and smoking history among the three groups. However, the history of allergic rhinitis and asthma, and the previous history of endoscopic surgery were different in these three groups (Table 1).

Table 1 Characteristics of the three groups: CRSwNP patients, CRSsNP patients, and control subjects
Full size table

In this study, 251 swab samples were collected, and the isolation rate for bacteria was 81.3% for the three groups (204 cases), with the lowest in the CRSsNP group (77.3%) and the highest in the CRSwNP group (88.4%) (Table 2). However, there was no significant difference in the isolation rate of microbial cultures among the three groups (P = 0.349).

Table 2 The isolation rate of microbial cultures in the three groups: CRSwNP patients, CRSsNP patients, and control subjects
Full size table

The difference in the number of bacteria in each group was not statistically significant (Table 3). A total of 22 positive strains were detected in 136 cases for the CRSwNP group. Among these cases, 33 (24.3%) were Coagulation-negative staphylococcus, 27 (19.9%) were Corynebacterium and 26 (19.1%) were Staphylococcus epidermidis. Twenty positive strains were detected in the CRSsNP group, including S. epidermidis (14 cases, 21.2%), Corynebacterium (14 cases, 21.2%), Coagulase-negative staphylococcus (12 cases, 18.2%) and S. aureus (9 cases, 13.6%). In the control group, 17 cases (34.7%) were detected with positive strains, among which 15 cases (30.6%) were Staphylococcus epidermal, 14 cases (28.6%) of Coagulase-negative staphylococcus and 7 cases (14.3%) of S. aureus. The positive strains were mainly Gram-positive aerobic and facultative anaerobic bacteria in these three groups, including 69.8% in the CRSwNP group, 58 (68.2%) in the CRSsNP group and 46 strains (74.2%) in the control group, followed by Gram-negative aerobic and facultative anaerobic bacteria in the three groups. Moreover, the isolation rates of Gram-negative obligate aerobic bacteria, anaerobic bacteria and fungi were very low. However, there was no statistically significant difference in the major detectable strains, except Citrobacter (Table 4).

Table 3 Numbers of different bacterial species in specimens from the three groups: CRSwNP patients, CRSsNP patients, and control subjects
Full size table
Table 4 Bacteria and fungi cultured from middle meatus specimens from the three groups: CRSwNP patients, CRSsNP patients, and control subjects
Full size table

Furthermore, when samples were categorized into subgroups according to the percentage of eosinophils, some bacterial species showed different rates in CRSwNP group (e.g., S. aureus, 3.3% in the subgroup with normal percentage of eosinophils, 17.2% in the subgroup with increased percentage of eosinophils, P = 0.011). Similarly, for S. epidermidis, the corresponding rates were 29.5 and 10.9%, respectively (P = 0.009).

Moreover, in the CRSwNP accompanied with or without asthma groups, the isolation rates of Corynebacterium (32.4% vs. 15.7%) and Pseudomonas aeruginosa (16.2% vs. 3.4%) were statistically different (P = 0.035 and P = 0.030, respectively). Similarly, patients with a history of ESS exhibited a lower isolation rate of Coagulase-negative staphylococcus (13.6%) and higher isolation rate of P. aeruginosa (16.2%) compared with patients without surgery (30.1 and 3.2%, respectively), and the differences were significant (P = 0.050 and P = 0.024, respectively). However, the other bacteria mentioned above showed no difference in the CRSsNP and control groups (Tables 5, 6, 7). In this study, we did not find the effect of smoking and allergic rhinitis factors on the distribution of microflora in each group.

Table 5 The effect of eosinophils on the distribution of microflora
Full size table
Table 6 The effect of asthma on the distribution of microflora
Full size table
Table 7 The effect of history of nasal surgery on the distribution of microflora
Full size table

Discussion

In this study, we found that the isolation rates for bacteria were not significantly different among the three study groups, and aerobic and facultative anaerobic bacteria were the main species. Specifically, Coagulase-negative staphylococcus and S. epidermidis were the most common species among the three groups. No significant differences in the isolation rates were observed among the three groups for the bacterial species with high isolation rates, whereas a high rate of Citrobacter with low isolation rates was observed in the CRSwNP group. Furthermore, when study samples were categorized into subgroups according to the percentage of eosinophils, some bacterial species showed different rates in the CRSwNP group. We did not find the effect of smoking and allergic rhinitis factors on the distribution of microflora in each group. These findings provide important insights into the mechanism underlying CRS and suggest the immunomodulatory effects of microflora in CRS.

The isolation rates for bacteria were not significantly different among the three study groups. The rates are generally consistent with that reported in the literature [23,24,25]. In this study, aerobic and facultative anaerobic bacteria were the main species, followed by Gram-negative aerobic or facultative anaerobic bacteria, strictly Gram-negative bacteria, and the isolation rates of anaerobic bacteria and fungi were very low. To date, different order and types of bacterial species have been reported. For example, Liu et al. [26] found that the aerobic bacteria were mainly Coagulase-negative staphylococcus, α-hemolytic streptococcus, whereas anaerobic bacteria were mainly Streptococcus and Streptomyces in 42 cases of adult maxillary sinusitis. Rombaux et al. [24] reported an order: Coagulase-negative staphylococcus, S. aureus, Streptococcus, other Gram-positive cocci, Haemophilus influenzae, non-fermented Gram-negative bacilli, Enterobacter, and anaerobic bacteria. In a study of 31 patients with or without nasal polyps, Niederfuhr et al. [25] found that the order of bacteria was Coagulase-negative staphylococcus, Bacteroides, S. aureus, α-hemolytic streptococcus. Liu et al. [27] reported that the common positive strains were Coagulase-negative staphylococcus, Corynebacterium, S. aureus, and H. influenzae. In our study, the most common bacterial species were Coagulase-negative staphylococcus, Corynebacterium, and S. epidermidis in the CRSwNP group; S. epidermidis, Corynebacterium, Coagulase-negative staphylococcus, and S. aureus in the CRSsNP group; S. epidermidis, Coagulase-negative staphylococcus, and S. aureus in the control group.

We found no significant differences in the rates for bacterial species with high isolation rates in middle meatus samples from CRSwNP patients, CRSsNP patients, and control subjects. This is consistent with results from prior studies [25, 27, 28]. Note that a high rate of Citrobacter (a bacterium with low isolation rate) was observed in the CRSwNP group. Citrobacter is a Gram-negative facultative anaerobic bacterium and is a common intestinal colonization bacterium. There are few studies on the colonization state of Citrobacter in nasal sinus and its role in the immune function of the host in CRSwNP patients, and more studies are needed in the future.

Recent studies have shown that there may be interactions between local microflora and the immune system. Smeekens et al. [29] have found that the number of normal bacteria (e.g., Corynebacterium) in patients with high IgE syndrome (HIES) was reduced, and the number of Gram-negative bacilli was increased (e.g., Pseudomonas). In peripheral blood mononuclear cell stimulation experiments, Pseudomonas can inhibit the cytotoxin against S. aureus, but the normal Corynebacterium does not have this inhibitory effect. Ba et al. [21] have found that the colonization of Gram-positive bacteria was more common in Chinese CRS patients with IL-5 positive nasal polyps, whereas the colonization of Gram-negative bacteria was more common in Chinese CRS patients with KCN (Key Cytokine-negative Nasal Polyps). Therefore, they suggested that the colonization of bacteria might be related to local IL-5 levels [21]. Ramakrishnan et al. [30] have found that the composition of microflora in CRS patients with complicated asthma and local purulent secretions showed some differences compared with patients with other phenotypes, suggesting that changes in microflora could disrupt the immune balance, and thus cause persistent inflammation status. Aurora et al. [28] have found that the blood leukocytes in CRS patients produced excessive IL-5 when they were exposed to symbiotic bacteria. Other studies also have observed that the composition of airway microflora may be associated with bronchial hyper responsiveness, increased eosinophil counts and total IgE levels [31, 32]. These findings suggest that changes in airway microflora may be associated with specific inflammatory processes.

In this study, we found that the isolation rate of S. aureus in the CRSwNP group was 11.0%, which was significantly lower than that reported in Europe [33,34,35], but was consistent with other studies in China [21, 36]. Furthermore, we found that the distributions of S. aureus and S. epidermidis showed opposite trend according to different EOS phenotypes in the peripheral blood. Compared with the subgroup with normal percentage of eosinophils, the isolation rate of S. aureus (3.3%) was lower than that in the subgroup with increased percentage of eosinophils (17.2%). However, the corresponding isolation rate of S. epidermidis decreased (29.5 and 10.9%, respectively). As there is a correlation between the phenotype of EOS in the peripheral blood and the local EOS phenotype of the nasal mucosa [37, 38], we postulate that similar results may be present according to the different local EOS phenotypes of the nasal mucosa. In the study on children with atopic dermatitis (Atopic dermatitis, AD), Laborel-Préneron et al. [39] found that S. aureus may promote inflammatory response through concomitant Th2 cell activation and Treg cell inhibition, and thereby promote the inflammatory response process of Th2 type, whereas normal flora such as S. epidermidis may counteract this effect by inducing skin DC cells to produce IL-10. It is unknown whether such a similar immune regulation mechanism exists in the mucosa of CRSwNP patients. That is, through the above mechanism, S. aureus induces eosinophilic inflammation, and promotes the inflammatory process of Th2 type, whereas S. epidermidis acts against this process, and promotes non-eosinophilic granulocyte inflammation. This hypothesis needs to be further verified.

In this study, we found that the isolation rate of P. aeruginosa and Corynebacterium was higher in CRSwNP patients with asthma than that in those without asthma. But there was no effect of asthma on the isolation rate of S. aureus in this group. Similarly, other researchers did not find any effect of asthma on the isolation rate of S. aureus based on Chinese populations [27, 29]. However, studies in Caucasian populations found that the isolation rate of S. aureus was the highest in CRSwNP patients with asthma [33,34,35]. We postulate that this disparity may be due to the differences in gene expression profiles and living conditions. Regarding the impact of asthma on the isolation rate of P. aeruginosa and Corynebacterium in CRSwNP patients, it is suggested that: (1) studies with large sample size are needed in the future; (2) consider the perspective of the interaction with microflora immune function.

This study found that the history of nasal surgery may have an impact on the isolation rate of nasal microbacteria. Prince et al. [40] reported that the isolation rate of nasal biofilm was 30.7% in patients with a prior history of functional endoscopic sinus surgery (FESS), which was higher than that in those without FESS (15.0%). In CRSwNP patients with surgical history, the isolation rate of P. aeruginosa increased, and the isolation rate of Coagulase-negative staphylococcus decreased. This suggests that surgical procedures may reduce the colonization of Coagulase-negative staphylococcus (normal flora) colonization, and increase the colonization of Pseudomonas, and thus affect the local microbial community status, and plays a role in local epithelial immune regulation.

Our study has some implications in clinical setting. For example, the isolation of microbacteria from the middle meatus during ESS becomes useful in the daily clinical practice because the corresponding results can help determine if the antibiotic therapy is needed. Furthermore, since the isolation rate of P. aeruginosa and Corynebacterium was higher in CRSwNP patients with asthma, and those patients also had high risk for recurrences of nasal polyps, the microbacterial profile might be useful in the future for early predicting more aggressive rhinosinusitis.

Conclusions

This study found no significant differences in the microbiological features in middle meatus samples from CRSwNP patients, CRSsNP patients, and control subjects, except for Citrobacter. The clinical significance needs further study. Furthermore, the distributions of S. aureus and S. epidermidis showed opposite trend according to different EOS phenotype in peripheral blood. The colonization rate of S. aureus was increased in subgroup with increased percentage of eosinophils, whereas the colonization rate of S. epidermidis was increased in subgroup with normal percentage of eosinophils. This finding suggests that S. aureus may promote eosinophilic inflammatory response, while S. epidermidis may promote non-eosinophilic inflammatory response. Our results suggest the immunomodulatory effects of microflora, and the non-infective factor may play a role in the pathogenesis of CRS.