PCR: Polymerase Chain Reaction; NTHi: Nontypeable Haemophilus influenzae ; AOM: Acute Otitis Media; DNA: Deoxyribonucleic Acid; MEE: Middle Ear Effusions; PBS: Phosphate Buffered Saline; Ct: Cycle Threshold

Acute otitis media (AOM) is one of the common bacterial infections in childhood [1], with nontypeable Haemophilus influenzae (NTHi) being a leading causative pathogen responsible for AOM [2-4]. The causative pathogens frequently colonize the nasopharynx and infect to the middle ear via the Eustachian tubes [5].

Although isolation of bacteria from MEEs has long been the gold standard for the etiologic diagnosis of the causative pathogens for AOM, only approximately 50%-60% of MEE samples from pediatric AOM are culture-positive for any of the major causative pathogens [2-4,6]. In 1993, we reported the first application of the polymerase chain reaction (PCR) to identifying NTHi in culture-negative MEEs from children with otitis media with effusion (OME) [7].

Gene amplification methods have increasingly been used in the etiological diagnosis of microbial infections in an effort to evaluate precise prevalence of pathogens present in middle ear effusions (MEEs) [8-14]. Although NTHi is a fastidious bacterium that is difficult to identify by conventional culture methods, gene amplification methods can reveal the existence of this bacteria in MEEs [7,15,16].

The recent advantages in molecular identification further emphasize the importance that bacterial density plays in the pathogenesis of infectious diseases [17-21]. Previously, we revealed that quantitative real-time PCR can be used to detect the increase in both the density of nasopharyngeal pneumococci and the ratio at which this pathogen can be detected during the onset of pediatric AOM [22].

In this study, we applied real-time PCR to quantify the amount of genomic NTHi DNA in MEEs and to evaluate the influence of the amount of bacteria in MEEs on the clinical outcome of AOM. The improvement of tympanic membrane abnormalities at second visits was significantly worse in cases with MEEs containing large amount of genomic NTHi DNA than it was in cases with MEEs containing small amount of genomic NTHi DNA. The amount of pathogens will be an important factor in predicting poor clinical outcomes of AOM.

Patients

Thirty-two children with severe AOM were evaluated in this study. They ranged in age from 6 months old to 8 years old (median 21.5 months old) and included 16 boys and 16 girls. All of the children had been patients at the Sugita ENT Clinics between July 2006 and June 2011. In addition, they all had acute illness lasting <48 hours, had no spontaneous ear drum perforations, and had not been treated with tympanostomy tubes. Myringotomy was performed at the first visit for treatment and approximately ≥70 μl of MEE was obtained by ATOMS tap (Luminas Co. Japan). Conventional bacterial culture was performed to identify the causative pathogens in the MEEs. The amount of genomic NTHi DNA in the MEEs was assessed by real-time PCR. Eighteen (56.3%) of the children were treated with cefditoren pivoxil at 18 mg/kg for 3 to 5 days. The remaining 14 were treated with various types of oral antibiotics for 3 to 5 days, with those antibiotics including amoxicillin clavulanate, cefcapen pivoxil, tosfloxiacin, and tebipenem pivoxil.

The diagnostic criteria and the assessment of severity for AOM were defined on the basis of the scoring system recommended by Japanese clinical practice guidelines for the diagnosis and management of AOM in children [23]. This scoring system makes use of such symptoms as fever and irritability and otalgia grades each of these parameters from 0 (none) and 2 (intermediate) to 4 (marked). The system also takes into consideration tympanic membrane changes including erythema, bulging, and otorrhea and grades each of these parameters from 0 (none) and 4 (intermediate) to 8 (marked) [23]. A total score summation of each parameter of 0, 1 to 8, 9 to 10, and 11 to 12 defined normal, mild AOM, moderate AOM and severe AOM, respectively.

Informed consent was obtained from the parents or guardians of the patients prior to collecting the samples. This study was approved by the ethical committee of Wakayama Medical University.

Bacterial cultures

The MEEs from the patients were examined by conventional bacterial cultures to determine the presence or absence of NTHi. Briefly, a small amount (approximately 10 to 20 μl) of each specimen was inoculated onto 5% sheep blood agar plates and chocolate agar plates (Nippon Becton Dickinson Company Ltd., Tokyo, Japan) immediately after the sample was collected. The plates were then incubated for 48 hours at 37°C in a humidified atmosphere supplemented with 5% CO₂. H. influenzae was identified by colony morphology, Gram-stained smear, and V and X factor requirements. S. pneumoniaewas identified by alpha-hemolysis and colony morphology, Gram-stained smear, optochin disk sensitivity, and bile solubility.

Real-time PCR

The relative densities of protein D gene (glpQ) of NTHi were determined by real-time PCR [24]. Briefly, MEEs (≥ 50 μl of volume) were suspended with ~1000 μl of phosphate buffered saline (PBS) to reduce the viscosity (2x to 10x dilution). Then, the genomic DNA was extracted from the MEE specimens using the QIAamp blood mini kit.

The real-time PCR assay was carried out in a final 25 μl reaction volume that include 12.5 μl TaqMan Universal Master Mix (Applied Biosystems, Foster City, Calif, USA) and 2.0 μl of DNA template in 96- well plates (Applied Biosystems). The primers and probe used in this study were forward primer: 5’-CTGGWGCAATGGCAGAAGTG-3’, reverse primer: 5’-TCTTTACGCACGGTGTAAGGATG-3’, and TaqMan probe: 5’-FAM-AATATGCCGATGGTGTTGGYCCAGGTTMGB- 3’. DNA was amplified with the Applied Biosystems 7500 Realtime PCR System (Applied Biosystems). After the initial denaturation (50°C for 2 minutes and 95°C for 10 minutes), the reaction was followed by 40 cycles of amplification of PCR products at 95°C for 15 seconds and at 60°C for 1 minute. Amplification data were analyzed by the instrument’s software (SDS, Applied Biosystems). RNase free water for negative control and DNA from respective serotypes for positive control serially diluted from 10¹ to 10⁷ CFU/ml were run for the standard curve in every test. The standard curve was generated based on the cycle threshold (Ct) values of the positive control.

The amount of protein D gene copies of NTHi in the samples were calculated based on the standard curve and adjusted based on the dilution ratio, as mentioned above. Samples with Ct values over 37 were defined as negative for real-time PCR. The sensitivity of the realtime PCR ranged between 0.022 pg/ml to 0.085 pg/ml. Finally, we established 10^-3 ng/ml as the cutoff value for the amount of the genomic NTHi DNA.

Statistical analysis

We performed statistical analysis using Prism 5 (GraphPad Software, Inc., La Jolla, CA, USA). The unpaired t-test was used to compare the amount of genomic NTHi DNA in the 32 children’s MEEs. Correlations were analyzed by the Pearson test. A p-value of <0.05 was considered statistically significant.

Distribution of the amount of H. influenzae protein D gene in MEEs

The 32 MEEs were divided into 12 culture-positive and 20 culturenegative specimens. With the real-time PCR, 20 out of 32 MEE samples contained protein D gene. When the amount of protein D gene in the MEEs from culture-positive and culture-negative specimens were determined by conventional culture methods, no differences were found between the culture-positive and culture-negative specimens (Figure 1). The 3 (25.0%) out of 12 MEE samples that were positive by conventional culture did not present the protein D gene. In contrast, the 11 (55.0%) out of 20 MEE samples that were negative by conventional culture presented the protein D gene.

Figure 1: Distribution of the amount of NTHi protein D genes in MEEs based on the conventional bacterial cultures. Open circle: NTHi-negative by conventional culture (n=17), Closed circle: NTHi-positive by conventional culture (n=12), Open square: S. pneumoniae-positive, NTHi-negative by conventional culture (n=3).

On the other hand, the 32 MEEs were also divided into 13 MEEs from children with prior antimicrobial treatment and 19 MEEs from children without prior antimicrobial treatments. No differences in the amount of protein D gene were found between the MEEs from children who had had or who had not had prior antimicrobial treatments (Figure 2). Seven (53.8%) out of 13 MEE samples from children with prior antimicrobial treatment presented the protein D gene. However, only two samples were NTHi-positive by conventional bacterial culture, with one being positive and the other being negative via real-time PCR. In contrast, 5 out of 19 MEE samples from children without prior antimicrobial treatment did not present the protein D gene and 3 of them were negative by conventional bacterial culture. Eight (28.6%) out of 14 MEEs containing genomic NTHi DNA were positive as determined by conventional bacterial culture.

Figure 2: Distribution of the amount of NTHi protein D genes in MEEs with or without the prior antimicrobial treatments. Open circle: NTHi-negative by conventional culture, Closed circle: NTHipositive by conventional culture, Open square: S. pneumoniae - positive, NTHi-negative by conventional culture.

Correlation between the amount of H. influenzae protein D gene in MEEs and clinical outcomes of AOM

At 3 to 5 days after their initial antimicrobial treatments and myringotomy, all 32 children returned for a second visit. 20 of them had NTHi DNA by real-time PCR. (Two of the 20 DNA-positive samples had pneumococci by conventional bacterial culture.) At these visits, the improvement ratios of the children’s tympanic membrane abnormalities were evaluated, with the improvement ratio being determined as the ratio of a) the reduction of tympanic membrane scores before and after initial treatments to b) the initial tympanic membrane score. The improvement ratio of tympanic membrane abnormalities of 20 MEE samples with protein D gene detected by realtime PCR was negatively correlated with the amount of protein D genes in the MEEs (p=0.004, r=-0.61) (Figure 3). When the all 32 MEE samples were evaluated, the amount of protein D genes in MEEs were significantly higher for cases in which the improvement ratio of tympanic membrane abnormalities at the second visit was below 50% than for cases in which the improvement ratio of tympanic membrane abnormalities at the second visit was above 50%. (In this analysis, MEEs with the cutoff value were included) (Figure 4).

Figure 3: Correlation between the amount of NTHi protein D genes in MEEs and the improvement of tympanic membrane abnormalities at 2nd visit. Open circle: NTHi-negative by conventional culture (n=9), Closed circle: NTHi-positive by conventional culture (n=9, # 3 cases overlapped with a similar value.), Open square: S. pneumoniae-positive, NTHi-negative by conventional culture (n=2).

Figure 4: Distribution of the amount of NTHi protein D genes in MEEs based on the improvement of tympanic membrane abnormalities at 2nd visit. Open circle: NTHi-negative by conventional culture, Closed circle: NTHi-positive by conventional culture, Open square: S. pneumoniae -positive, NTHi-negative by conventional culture.

Acute otitis media is a leading reason that oral antibiotics are prescribed during childhood [25,26]. The Japanese clinical practice guideline recommends evaluating the severity of AOM cases when diagnosing and managing AOM in children [23]. However, the effects of antimicrobial treatments have limitations. Cases of AOM caused by NTHi sometimes persist after initial antimicrobial treatments [27-30]. In addition, the high clinical failure rates following antibiotic treatment are related to the microbial nature of AOM [31]. On the other hand, when Stenfors et al. reported on their use of acridine orange stain to quantify the bacteria in MEEs in the early 1990s, they suggested the importance of quantitative evaluation of middle ear pathogens [32-34]. However, little effort has been focused on determining and using bacterial density as a factor for characterizing the clinical outcomes of AOM.

In this study, we found that the amount of genomic NTHi DNA was not different between culture-positive and culture-negative samples. In addition, this study did not find any association between previous antibiotic treatment and the amount of NTHi in the MEEs. On the other hand, previous antibiotic treatment did reduce the ability of conventional bacterial cultures to detect NTHi. In contrast to the successful application of molecular procedures to identify pathogens in MEEs, there are still considerable controversies over whether or not the bacterial DNA detected by PCR implies the existence of metabolically active bacteria.

Post et al. used a chinchilla AOM model to reveal that DNA from dead bacteria does not remain detectable in MEEs [35,36]. Palmu et al. also showed that the PCR-positive rate for MEE samples decreases with time in cases of pneumococcal AOM [37]. Detection of mRNA by RT-PCR further suggested that PCR-positive but culture-negative MEEs would be caused by metabolically active NTHi [38,39]. Further, MEEs may have nuclease-inhibiting activity, and both DNA and mRNA can persist longer in MEEs. Thus, the amount of NTHi genomic DNA in MEEs will reflect the number of NTHi cells infecting the middle ear cavity while real-time PCR can detect both dead and viable NTHi.

The aim of this study was to apply real-time PCR to quantify the amount of genomic NTHi DNA in MEEs and to evaluate the influence of the amount of bacteria in MEEs on the clinical outcome of AOM. In this study, the amount of genomic NTHi DNA in MEEs was correlated with the clinical improvement of severe cases of AOM. Treatment failures and relapses of AOM are caused by several microbiological factors such as antibiotic resistance, viral-bacterial coinfection, biofilm formation, and invasion of bacteria into epithelial cells [40-42]. The presence of bacterial biofilms on mucosal membranes has been proposed as an explanation for cases having positive PCR results yet negative culture results [43,44]. In a biofilm, the bacteria may form a complex bacterial community attached to middle ear epithelium and covered by a protective polysaccharide matrix. These biofilms are resistant to antimicrobials and immune defense mechanisms. While they are in the biofilm phenotype, bacteria interact less with the environment and are more difficult to isolate in culture, but they may be able to survive in the middle ear for long survive.

Concerning the ability to predict the clinical outcomes of AOM, Yano et al. reported that the improvement rate of AOM based on the severity score at 1 week after the initial examination was significantly associated with the time to recovery [45]. While their findings at 1 week after the initial treatment are important to predicting the outcome of AOM cases, their findings were defined relatively late to predict clinical outcomes of AOM in a usual clinical setting.

The study reported here has not revealed the relationship between the amount of bacteria in MEEs and the severity of AOM. However, our unpublished study showed that the amount of bacteria in MEEs was related to the severity of AOM. Nonetheless, our current findings suggest that the severity of tympanic membrane abnormalities as evaluated using the scoring system proposed by Japanese AOM guideline will be better understanding the amount of NTHi in middle ear and that resulted in the poor improvement of AOM.