Conjunctival sac is colonized by a diverse range of microorganisms which constitute the normal ocular flora or potential pathogens. However, these microorganisms are capable of causing infections of conjunctiva and cornea when changes happen to microenvironment of the ocular surface or systemic condition. Especially when the globe of the eye is breached by trauma or surgery, these floras could be the predisposing factors of the endophthalmitis (Armstrong, 2000; Deorukhkar et al., 2012). Due to multiple environments as well as different ethnic groups and patterns of administration, the global epidemiology indicates that the spectrum of conjunctival sacs pathogens and resistance patterns vary between geographic locations (Bertino, 2009; Passos et al., 2010). Moreover, reports of microbial isolates from either diseased or preoperative ocular tissues varied widely according to published studies (Khristov, 1956; Sharma et al., 2013).

This study was carried out in Chinese population of substantial size, which seeked to define the incidence and antimicrobial sensitivity profiles of normal conjunctival bacterial flora in the central area of China. Since eye care professionals regularly provide antibiotic prophylaxis to their patients following ocular trauma/injury, such data could be used to guide the selection of appropriate prophylactic agents. Thus, our study could be useful in the prevention of ocular infections and in choosing the initial empirical therapy prior to final laboratory results.

A total of 8,224 conjunctival samples collected from 4,266 males (51.9%) and 3,958 females (48.1%) were divided into 5 groups (Table 1). The median age of the patients was 37.0 ± 28.2 years and the age difference between males and females was significant (χ² = 7.107, P = 0.008). The isolation rate in the age groups differed significantly (χ² = 57.346, P < 0.01), namely, 29.0% (506/1,743) in 0–6 years group, 25.5% (377/1,478) in adolescence group, 21.7% (230/1,059) in young group, 19.4% (423/20,184) in the 41–65 years group, and 25.8% (454/1,760) in the elderly group. Age was an influencing factor on bacterial isolation in the patients with different gender. For example, the incidence of bacterial isolation in males (30.7%) was significantly higher than that in females (22.1%) in elderly group (χ² = 16.674, P < 0.01). As shown in Figure 1, the incidence of bacterial isolation showed a trend of decreasing at first, and then climbed up along with age. The middle aged group of 41–65 years presented the lowest rate of bacterial isolation (19.4%), while the highest isolation rate (83.1%) was found in patients in the age group of 0–6 years. Furthermore, in every age group, the incidence of bacterial isolation was higher in men than that in women.

Of the 1,990 isolated strains, 920 (46.2%) strains were only gram stained without performing specific strain identification and susceptibility test for the following reasons: 887 strains were gram-positive corynebacterium, and the majority were normal flora of human skin and mucosa, which needed to be further identified only in special circumstances (Wang et al., 2015); 33 strains of gram-positive cocci were also not identified. In spite of these unclear isolations above, we eventually isolated 1,070 strains and 24 types of flora. The number of isolation types of flora was 19 in children group, 10 in adolescence group, 10 in young group, 8 in middle aged, and 12 in elderly group, respectively. The most commonly isolated bacteria was S. epidermidis (39.7%, 790/1,990), surpassing S. pneumoniae (4.5%, 89/1,990) which was followed by S. aureus (2.7%, 54/1,990) and Moraxella catarrhalis (2.2%, 44/1,990). The isolation rates exhibited statistically significant difference in each age group (Table 2).

The susceptibility profiles of different isolates also differed with type. CONS was the most frequently isolated flora from the conjunctival sacs. Most CONS were resistant to penicillin (90.8%), followed by azithromycin (85.8%), erythromycin (84.4%), oxacillin (63.4%), and clindamycin (57.9%). Except moxifloxacin (4.0%), the resistance rates for CONS to quinolones were observed between 23.1 and 38.6%. Besides, CONS were susceptible well to a spectrum of antibacterial agents such as rifampicin (3.5%), minocycline (1.0%), gentamycin (6.9%), and doxycycline (1.2%). As shown in Figure 2, we discovered that 2 species of CONS were resistant to vancomycin, and no strain of CONS was resistant to linezolid.

By contrast with CONS, S. aureus demonstrated lower resistance rates. It performed like that: 76.8% of isolating CONS were resistant to oxacillin, while in S. aureus the rate was 14.8%. To co-trimoxazole, the resistance rates of CONS and S. aureus were 63.4 and 13.0%, respectively. And the resistance rates to azithromycin and erythromycin were 85.8 vs.72.7%, and 84.4 vs. 72.2%, respectively. All the differences were statistically significant. Comparing to CONS with resistance rate 4.0–38.6%, S. aureus showed lower resistance rate with 0.0–9.3% to quinolone. There was no resistant strain to vancomycin, linezolid, rifampicin, or minocycline among S. aureus (Figure 3).

As demonstrated in Figure 4, Streptococcus was sensitive to vancomycin, linezolid, and levofloxacin, whereas highly resistant to erythromycin (98.8%), clindamycin (96.3%), and tetracycline (82.7%).

The support of conjunctival sac on bacteria may result from the direct contact with the outside environment and the connection to the adjacent skin, and the different results of culture are greatly attributed to various factors such as environment, age, administration of antimicrobial agents, etc. (Barfoed, 1956; Khristov, 1956; Sharma et al., 2013). The use of antimicrobial agents contributes to the emergency of the new drug resistant strains (Leibovitch et al., 2005; Recchia et al., 2005), therefore, the pathogens of ocular infectious diseases are more attractive to the clinical researchers. But only few investigations have focused on the microflora of normal healthy conjunctival sacs. Our study analyzed the results of culture and antimicrobial susceptibility testing of 1,070 types of normal conjunctival flora, expecting to elaborate the diffusion and antimicrobial susceptibility of regional normal conjunctival flora, thereby making a contribution to the clinical recommendations for the prescription.

In this study, the overall incidence of isolated bacteria was 24.2%, lower than the published results of 36.0–81.7% (Ta et al., 2009; Sharma et al., 2013; Muluye et al., 2014). Previous publications reported that certain local anesthetic eye drops had strong antimicrobial effects, which were dose-related and related to whether the drops contained preservatives (Mullin and Rubinfeld, 1997; Dantas et al., 2000; Pelosini et al., 2009). Compared to published results, we detected a relatively lower incidence of conjunctival flora in present study, and the use of local anesthetic might be contributory. Twenty-four species of conjunctival micro-organism were isolated altogether, and the difference of bacterial isolation among various age groups was statistically significant. The isolated rate showed a trend of decreasing at first and then climbed up along with age. The middle aged group of 41–65 years presented the lowest rate of bacterial isolation (19.4%). Further investigation showed that in every age group, the incidence of bacterial isolation in men was higher than that in women, especially in elderly group (>65 years), and the difference was significant (Singer et al., 1988; Li and Smith, 2002).

The most commonly isolated micro-organism was S. epidermidis (39.7%), which was followed by S. pneumoniae (4.5%) and S. aureus (2.7%). S. epidermidis was the predominant isolate, that was consistent with previously reported trends (Ta et al., 2009; Sharma et al., 2013; Muluye et al., 2014). The majority (83.1%) of the bacterial isolates were from patients in the age range of 0–6 years. Gram negative isolates were found in small numbers only, with M. catarrhalis being most common (2.2%), and we also isolated other strains of gram negative rods such as P. aeruginosa, Philo acinetobacter, Klebsiella pneumoniae, and Proteus mirabilis.

It was thought that normal ocular flora could be non-pathogenic or occasionally pathogenic. However, the pathogens of some bacterial endophthalmitis, bacterial corneal ulcers, blepharitis, conjunctivitis, and other ocular infection diseases turned out to consistent with conjunctival isolated bacteria, and the S. epidermidis has become the predisposing pathogen (Recchia et al., 2005; Carreras, 2012; Sharma et al., 2013). Although, the preoperational topical antibiotics can hardly make the conjunctival sacs sterile, they can effectively reduce the isolating bacteria (Mshangila et al., 2013). The susceptibility of each antibiotic to different bacteria is not identical, and the regional reported drug resistance varies widely due to different environment and the use of antibiotics (Carreras, 2012; Mshangila et al., 2013; Muluye et al., 2014). The overall incidence of isolated bacteria was 24.2%, and we found higher resistance rates to penicillic, macrolides, clindamycin and sulfonamides by CONS, ranging from 57.9 to 90.8%. With a relatively high susceptibility to moxifloxacin, the CONS was moderately susceptible to other fluoroquinolones, and highly susceptible to vancomycin, linezolid, rifampin, tetracyclines, and aminoglycosides. Contrasting to CONS, the general resistance rate of S. aureus was lower. Streptococcus was susceptible well to the majority of antimicrobial agents, while highly resistant to macrolides and tetracyclines with the rates >80%.

Evolving bacterial resistance represents one of the most serious global public health problems, and overcoming this problem has become a great challenge. Due to different ethnic group, environment and antimicrobial therapy, the distribution and resistant profiles of conjunctival bacteria vary significantly from area to area. Therefore, the investigation in these subjects can be clinically useful in the primary empirical antimicrobial strategy before knowing the laboratory results.

HT and LeiL conceived the study. LeiL and MC performed the laboratory experiments. JW and HZ contributed to the statistical analysis. HT wrote the draft of the paper. HT and LeL reviewed the manuscript. All authors have read and approved the final manuscript.