Incidence Survey of Acute Otitis Media in Children in Sado Island, Japan—Sado Otitis Media Study (SADOMS)

Background Acute otitis media (AOM) is one of the most common forms of bacterial infection and cause for clinic visits in children. The incidence of AOM was 0.9–1.2 episodes per person-year during the first 2 years of life in previous reports conducted before 2000. The aim of this study was to 1) evaluate the latest AOM incidence in pediatric outpatients and 2) identify the bacterial pathogens from these patients and ascertain their serotypes and resistance. Methods The study was conducted in a closed population, involving all pediatricians and otolaryngologists in Sado Island allowing accurate determination of AOM incidence. In each month, one week was assigned as “surveillance week”, and all outpatients with acute illness aged 0–18 years examined during the surveillance weeks were enrolled. AOM was diagnosed on the basis of otoscopic findings and clinical symptoms were recorded. Specimens were collected from the nasopharynx or middle ear cavity of AOM patients and examined for bacteria. Antimicrobial susceptibilities, serotypes, and molecular typing for resistance were determined among Streptococcus pneumoniae and Haemophilus influenzae. Results In total, 8,283 clinic visits were conducted, and 354 episodes (4.3%, 95% CI: 3.9–4.7%) among 312 children were diagnosed as AOM. The incidence of AOM was highest in children of 1 year of age (0.54 episodes/child/year, 95% CI: 0.44–0.64). Serotype coverage of 7- and 13-valent pneumococcal conjugate vaccines in this study were 38.0% (95% CI: 29.3–47.3) and 62.8% (95% CI: 53.6–71.4), respectively. Of 122 H.influenzae isolates available for typing, 120 were nontypeable and 2 were type b. A high proportion of S. pneumoniae isolates (46%) showed resistance to penicillin. Approximately half of H. influenzae isolates had genetic markers for beta-lactamase-negative ampicillin-resistance. Conclusions Approximately 4–5% of pediatric outpatients, even without AOM-related symptoms, had AOM in our study. Pediatricians as well as otolaryngologists should check the tympanic membrane findings of all pediatric outpatients.


Introduction
Acute otitis media (AOM) is one of the most common forms of bacterial infection and cause for clinic visits in children, with 70% of children experiencing at least one episode by age 3 [1,2]. The incidence of AOM were reported as 1.1-1.2 episodes per personyear during the first 2 years of life in the United States (conducted between 1975 and 1984) [2], 1.16-1.24 episodes per person-year from 6.5 to 24 months of age in Finland (conducted between 1995 and 1999) [3], and 0.9 episodes per person-year during the first 2 years of life in Bangladesh (conducted between 1993 and 1996) [4]. However, these rates were estimated during the 20 th century, and there have been time-dependent changes in the current conditions surrounding children, such as vaccines, antimicrobials, day-care use, living environment, and allergies.
Streptococcus pneumoniae, nontypeable Haemophilus influenzae (NTHi), and Moraxella catarrhalis, which are members of the commensal flora of the nasopharynx, are the 3 leading causes of AOM. The frequency of nasopharyngeal colonization by these pathogenic bacteria is strongly associated with AOM [5,6].
There have been few studies on the epidemiology (incidence, severity, and etiology) of AOM in Japan. According to the subcommittee regarding AOM guidelines, the exact frequency of AOM occurrence in Japan is unknown [7]. The aim of this study was to 1) evaluate AOM proportion in pediatric outpatients and incidence in Sado Island, Niigata, Japan and 2) identify the bacterial pathogens from these patients and ascertain their resistance.

Study Area and Population
The study, Sado Otitis Media Study (SADOMS), was conducted in Sado Island, which lies 60 km off the Japanese mainland, between January 2010 and December 2011. In Japan, H. influenzae type b (Hib) conjugate vaccine and 7-valent pneumococcal conjugate vaccine (PCV7) were introduced for voluntary immunization in December 2008 and February 2010, respectively. However, they have not been implemented as part of a routine vaccination program in Japan, and they have only been provided free of charge since January 2011 (See Table 1). SADOMS phase I was undertaken between January 2010 and December 2011 (the PCV7-introductory era), phase II will be undertaken between January 2012 and December 2013 (early phase of the post-PCV7 era), and phase III will be undertaken between January 2014 and December 2015 (late phase of the post-PCV7 era). At the time when the study was started (January 2010), the population of the island was 64,120, with 9,388 (14.6%) aged 0-18 years. Children aged less than five years constituted 3.4% (2,167) of the total population. Here, we report on SADOMS phase I.

Surveillance
The second week of each month was assigned as a surveillance week for observational surveillance. When the second week had a national holiday, the first or third week was assigned as a surveillance week. All outpatients with acute illness aged 0-18 years examined during the surveillance weeks were enrolled in this study. Acute illness was defined as the presence of one or more of the following symptoms: fever, cough, rhinorrhea, fatigue, nausea, vomiting, diarrhea, otalgia, exanthema, and irritability. Because symptoms of respiratory tract infections among children can be non-specific, our definition of acute illness could include nonrespiratory tract infections, such as gastrointestinal infection, urinary tract infection, and acute phase of chronic diseases. Children visiting a doctor for health checkups, vaccination, or treatment of chronic diseases were excluded from this study. No outpatient could be enrolled twice in the same surveillance week, while each outpatient could be enrolled every surveillance week. Each pediatric outpatient was seen by a doctor belonging to the SADOMS study group in Sado Island. The group included all pediatricians (6 doctors) and all otolaryngologists (3 doctors) in Sado Island. We covered 97.8% (10,310/10,538 visits) of the pediatric patients who visited hospitals/clinics due to acute illness during their first 3 years in Sado Island [8]. Although some outpatients .3 years old are not followed by the study group doctors, the referral system in Sado Island functions such that children with AOM regardless of age are referred to the 6 pediatricians or the 3 otolaryngologists. Therefore, it is likely that almost all AOM patients were enrolled in this study.
AOM was diagnosed on the basis of tympanic membrane findings; 1) the presence of middle ear effusion, and 2) signs of middle-ear inflammation including redness and bulging [9,10]. Also, we recorded a history of acute onset of signs and symptoms, such as fever, irritability, and otalgia (but, they are non-specific). AOM severity was diagnosed as mild, moderate, or severe, based on the scoring system included in the 2009 Japan AOM clinical guidelines [7]. In addition, an outpatient who was enrolled in the surveillance was counted as having AOM, if he/she was diagnosed with AOM within 10 days from his/her clinic visit. The reason is that most AOM cases cannot be diagnosed at first visit soon after developing the first symptoms of acute illness.
A tympanogram is a useful adjunct for the diagnosis of AOM in doubtful cases and has a 90% sensitivity to predict middle ear effusion [11]. However, a tympanogram is not always feasible in pediatric departments in Japan. Thus, our study pediatricians relied instead on otoscopic findings. To standardize the otoscopic skills, all 10 doctors who are experienced clinicians, were trained again prior to the study and used the same atlases/texts for AOM diagnosis based on the guidelines [9]. As part of our study protocol, pediatricians consulted otolaryngologists when the pediatrician's diagnosis of AOM was doubtful. The indications for a myringotomy were 1) tympanic membrane bulging or middle ear fluid, and 2) severe ear pain. An otolaryngologist determined the necessity for myringotomy based on AOM severity and patient's age.
The subjects' medical information was obtained from their patient's case records and by questionnaires. Information included i) the past history of AOM, and ii) vaccination status with Hib and PCV7 vaccines.
The AOM proportion for all outpatients with acute illness was calculated in each surveillance week. In addition, clinic visit-based incidence of AOM by age (episodes/child/year) was calculated from the census of Sado city. The number of AOM patients during the study period was estimated by multiplying the observed episodes by 104/24 (total number of weeks in 2 years study period/number of weeks of observation). Dividing by 2, we obtained the annual number of AOM episodes. Then, dividing by each age population, the clinic visit-based incidence of AOM (episodes/child/year) was calculated. We assumed that each ''surveillance week'' represented the month. Otitis prone was defined as 3 separate episodes in 6 months or 4 episodes in 1 year.

Bacterial Cultures
Specimens were collected from the nasopharynx or middle ear cavity and examined for bacteria. The middle ear fluid samples were obtained by myringotomy or from ear discharge.
Except for myringotomy samples, these cultures were obtained from AOM patients using a sterile cotton-tipped aluminum swab (Seed Swab No. 2; Eiken Chemical Co. Ltd., Tokyo, Japan). Swabs were cultured immediately on chocolate and sheep blood agar plates and incubated at 36uC in 5% vol/vol CO 2 for 18-24 hours. S. pneumoniae strains were identified by colony morphology, susceptibility to optochin, and bile solubility. H. influenzae strains were identified by colony morphology and the use of an automated diagnostic system (HNID panel with the MicroScan Walk Away-96; Dade Behring Inc., West Sacramento, CA). Then, the pathogens were confirmed as H. influenzae isolates by a PCR method identifying the 16S ribosomal DNA and IgA protease gene [12]. M. catarrhalis strains were identified by colony morphology and the presence of butyrate esterase. The predominant colony type on each plate was selected for this study.
All isolates of S. pneumoniae and H. influenzae were stored at 280uC until further analysis.

Antibiotic Susceptibility Testing
Antimicrobial susceptibilities were determined using commercially available MIC panels (MICroFAST 3J for S. pneumoniae and 4J for H. influenzae; Dade Behring Inc., Deerfield, IL, USA). MICs were interpreted as susceptible, intermediate, or resistant according to the guidelines set by the Clinical and Laboratory Standards Institute (CLSI) [13]. The following antimicrobial agents were tested: penicillin G (PEN), erythromycin, and clindamycin for S.  pneumoniae; and ampicillin (AMP) and amoxicillin-clavulanic acid (AMC) for H. influenzae.

Serotyping
The isolates of H. influenzae and S. pneumoniae were serotyped with a commercially available slide agglutination kit (Denka Seiken Co., Tokyo, Japan), and by the capsular swelling method (Quellung reaction) using commercially available antisera (Statens Serum Institutes, Copenhagen, Denmark). Isolates that showed negative reactions for all pooled sera were considered to be nontypeable.

Molecular Typing for Resistance
Molecular typing for resistance was determined by PCR. Bacterial DNA was extracted from the isolates using a High Pure PCR Template Preparation Kit (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions. Macrolide-lincosamide-streptogramin B (MLS B ) resistance genes (ermB and mefA/E), and mutations for penicillin resistance (pbp 1a, pbp 2x, and pbp 2b mutations) in S. pneumoniae were identified by PCR [14][15][16]. S. pneumoniae isolates were classified into 3 genotypes by the presence of pbp 1a, pbp 2x, and pbp 2b mutations: penicillin susceptible S. pneumoniae (gPSSP, no mutation), penicillin intermediate S. pneumoniae (gPISP, with 1 or 2 mutations), and penicillin resistant S. pneumoniae (gPRSP, with all mutations).

Data Analysis
The vaccination coverage by age among the entire Sado population and AOM patients, and the rates of bacterial carriage for each level of severity were compared. These analyses were performed by chi-square test and Fisher's exact test. Adjusted standardized residuals were calculated, when the rate of bacterial carriage was significantly different between groups based on AOM severity. P-values ,0.05 were considered significant. All statistical calculations were performed with PASW Statistics 18.0 for Windows (SPSS Inc., Chicago, IL, USA).

Ethics Statement
Because this study was an observational study and sampling of bacteria was a routine work in AOM treatment, verbal informed consent was obtained from the guardians of all children with AOM when the questionnaire was filled out. The bacteria samples and data sheets were anonymized. This study protocol including consent procedure was approved by the Ethics Committee in Sado General Hospital.

Results
In total, 8,283 (male 4,432 and female 3,851) clinic visits for acute illness were conducted, at which 354 AOM episodes (4.3%, 95% CI: 3.9-4.7%) were detected (Figure 1). Considering recurrent episodes, 312 children were diagnosed as AOM at least 1 time during study period. AOM proportion in outpatients were 5.5% (95% CI: 4.9-6.1%) among patients aged 0-5 years, and 1.8% (95% CI: 1.4-2.4%) among patients aged 6-18 years. Of 8,283 outpatients, pediatricians consulted otolaryngologists for ,100 children with diagnosed AOM and doubtful AOM. During this study, the number of clinic visits in each surveillance week ranged from 269 (in July 2011) to 402 (in May 2010 and November 2010). The number and proportion of AOM episodes among outpatients ranged from 5 (1.4%) in September 2011 to 25 (6.7%) in April 2011. The analysis for seasonality was not done, because we judged that 2 years surveillance was not enough for such an analysis.
Of 312 children with AOM, 162 (51.9%) were the first episode and 146 (46.8%) had recurrent episodes, including 21 (6.7%) otitis prone children, at the time they were diagnosed with AOM in this study. Past AOM histories in 4 AOM patients were not available. The clinic visit-based incidence of AOM was highest in children of 1 year of age (0.54 episodes/child/year, 95% CI: 0.44-0.64) and tended to decrease in older children (Table 1). Seventy and 143 were diagnosed as severe and non-severe AOM, respectively, among AOM patients aged 0-2 years. On the other hand, 16 and 125 were diagnosed as severe and non-severe AOM, respectively, among AOM patients aged .3 years. Younger children were more likely to be diagnosed with severe AOM (p,0.001).
Most patients with AOM (83.9%, 260/310) had at least 1 AOM pathogen and nearly half of the patients (43.9%, 136/310) had 2 or more AOM pathogens. As shown in Table 2, there were significant differences in the carriage rates of these nasopharyngeal pathogens between severe, moderate, and mild AOM. H. influenzae isolates were more likely to be detected in moderate AOM cases, while M. catarrhalis were more likely to be detected in severe cases.

Discussion
We conducted a hospital-based AOM surveillance which included almost all AOM cases in Sado Island. We observed that 4.3% (95% CI: 3.9-4.7) of pediatric outpatients (including those with and without the AOM-related symptoms mentioned above) had AOM in our study, indicating that pediatricians should check the tympanic membrane findings of all pediatric outpatients regardless of the presence of symptoms. A history of acute onset of signs and symptoms is controversial for diagnosis of AOM [9,10]. Shaikh, et al. suggest that tympanic membrane bulging is sufficient to diagnose AOM [10].
In addition, we report the clinic visit-based incidence of AOM in an area with limited movement of people (particularly of children) to and from other areas. AOM incidence was highest in children of 1 year of age 0.54 episodes/child/year (95% CI: 0.44-0.64) and tended to decrease in older children. A limitation of our study design is that children with AOM who did not visit hospitals/clinics would be missed. However, our main purpose was to identify the incidence of AOM in all outpatients. It is important for pediatricians to know the AOM frequency and to pay attention to AOM patients in the actual exam room.
AOM patients aged 0-1 year showed lower vaccination coverage for PCV7, compared to all children aged 0-1 year in Sado Island (Table 1). A possible explanation is that these children without PCV7 vaccination were more likely to develop AOM. However, only 38.0% of pneumococcal isolates were serotypes that are included in PCV7 in this study. Therefore, we speculate that missing vaccination opportunities caused by illness, such as recurrent infections including AOM, is also an important factor. The observation that younger AOM children had the low coverage of Hib vaccination provides support for the latter explanation, because Hib vaccination has no (or very little) effect to prevent AOM.
Some of the limitations in this study include subject enrollment, sample isolation, and the number of isolates available for genetic characterization. During the time that we enrolled 269-402 patients with acute illness, over 300 patients also visited our hospitals/clinics with chronic diseases, for vaccination, and for follow-up examination in each surveillance week. We decided to use ''surveillance week'' system because the trends of infectious diseases may gradually change in Sado Island. Over 90% of the cultures evaluated in this study were isolated from the nasopharynx, which included only colonization pathogens. Thus, we could not determine whether these pathogens were causative agents or not. However, there are differences in serotype proportion of S. pneumoniae between healthy children and AOM patients.
Approximately 4-5% of pediatric outpatients, even without AOM-related symptoms, had AOM in our study. Pediatricians as well as otolaryngologists should check the tympanic membrane findings of all pediatric outpatients. Pediatric AOM caused by antimicrobial resistant S. pneumoniae and H. influenzae are still a major concern in Japan. PCV13 as well as vaccines for NTHi could prevent the infection and thus AOM.