Introduction
Otitis media with effusion is a common medical problem, and a leading cause of hearing impairment in children presenting to otolaryngology clinics. It is defined as the presence of fluid in the middle ear, without any signs or symptoms of acute ear infection.Reference Riding, Bluestone, Michaels, Cantekin, Doyle and Poziviak1 Otitis media with effusion was previously thought to be a noninfectious disease secondary to eustachian tube dysfunction. However, several microbiological studies have now shown the presence of bacteria in the middle-ear fluid of patients with this condition.Reference Riding, Bluestone, Michaels, Cantekin, Doyle and Poziviak1–Reference Pereira, Pereira, Cantarelli and Costa5
The inappropriate and excessive use of antibiotics in the treatment of acute otitis media and other respiratory tract infections has raised concern about an increased prevalence of pathogens resistant to commonly used paediatric antibiotics.Reference Pelton6
Several studies in different countries have shown that Streptococcus pneumoniae and Haemophilus influenzae are among the most frequent causative organisms of acute otitis media, followed by Moraxella catarrhalis.Reference Pereira, Pereira, Cantarelli and Costa5, Reference Li, Chiu, Hsu, Lee, Hwang and Huang7, Reference Martinez, Macios and Palau8 Furthermore, the bacteriology of acute otitis media has been shown to change over time.Reference Kilpi, Herva, Kaijalainen, Syrjanem and Takala9, Reference Ruohola, Meurman, Nikkari, Skottman, Heikkinen and Ruskanen10Streptococcus pneumoniae used to be the most frequent bacteria associated with acute otitis media in the early twentieth century; however, nowadays M catarrhalis and H influenzae are becoming more frequent.Reference Kilpi, Herva, Kaijalainen, Syrjanem and Takala9
In a 2001 study of acute otitis media, the three major pathogens found in 329 children aged two to 24 months were S pneumoniae (26 per cent), M catarrhalis (23 per cent) and H influenzae (23 per cent).Reference Kilpi, Herva, Kaijalainen, Syrjanem and Takala9 In addition, M catarrhalis infections were shown to peak at a very young age (six months), while S pneumoniae and H influenzae infections peaked at 12 and 20 months, respectively.Reference Kilpi, Herva, Kaijalainen, Syrjanem and Takala9
Ruohola et al. found that new bacteria appeared in the middle ear more commonly during the follow-up period than at initial detection; on the first day of study, the above three common pathogens were detected in 89 per cent of 75 children.Reference Ruohola, Meurman, Nikkari, Skottman, Heikkinen and Ruskanen10Moraxella catarrhalis seemed to persist the longest in the middle ear.
In 1993, El-Shamy examined the bacteriological profile of otitis media in Egyptian children, and found that 50 per cent of all effusions yielded bacterial growth.Reference El-Shamy11Haemophilus influenzae was the commonest isolated organism, followed by M catarrhalis and S pneumoniae.
In 2005, Barkai et al. raised concerns about antimicrobial drug resistance in S pneumoniae found in children with acute otitis media who were aged less than five years and lived in southern Israel.Reference Barkai, Greenberg, Givon-Lavi, Dreifuss, Vardy and Dagan12
A 2007 review of persistent and recurrent acute otitis media occurring in 1077 children enrolled between 1995 and 2000 found that the most common cause of early recurrent acute infection in antibiotic-treated children was persistent S pneumoniae colonisation in the nasopharynx.Reference Leibovitz13 The author suggested that initial antibiotic treatment of acute otitis media should also target nasopharyngeal colonisation.
Due to the worldwide variation in the profile of middle-ear pathogens, the current study was conducted to evaluate the epidemiology and resistance profiles of bacteria implicated in otitis media with effusion in Lebanese children. The study also assessed the correlation between bacterial resistance profiles and other factors, such as smoking exposure at home and recurrent upper respiratory tract infection.
Materials and methods
This study was conducted prospectively in a single university hospital between January 2009 and April 2010. We enrolled 62 children (107 ears), who underwent myringotomy with tympanostomy tube placement for persistent otitis media with effusion.
The inclusion criteria were paediatric patients aged less than 13 years, who had otitis media with effusion diagnosed by pneumatic otoscopy, with resultant hearing impairment with or without language delay, and who had suffered at least six weeks' illness without any sign of resolution or improvement despite medical treatment (nasal steroids with or without reflux management).
The diagnostic criteria for otitis media with effusion included tympanic membrane opacification, colour change, decreased mobility and increased vascularity.
We excluded from the study any patients who presented with acute otitis media and other upper respiratory infections at the time of surgery, as well as those who had taken antibiotics within the past seven days prior to surgery.
The duration of otitis media with effusion was estimated based on the duration of hearing impairment noted by the parents, and also, more accurately, via follow up with pneumatic otoscopy. A detailed medical history was taken for each patient, focussing on relevant aspects such as age, sex, duration of hearing impairment, history of upper respiratory tract infection, smoking exposure at home and prior antibiotic use.
Surgery was performed under microscopy. After cleansing and antisepsis of the external auditory canal with 70 per cent alcohol, a myringotomy was made in the anteroinferior quadrant of the tympanic membrane.
The effusion was collected with a number five aspirator connected to a Polymed mucus extractor (POLY MEDICURE LTD, New Delhi, India), using aseptic technique. Effusion fluid was sent for direct culture no longer than 15 minutes after collection.
For direct culture, the effusion fluid was seeded onto plates containing sheep blood agar and chocolate agar (enriched with PolyViteX; BIOMERIEUX, France SA), and incubated in aerobic conditions under 5–10 per cent CO2 for 48 hours at 37°C. The resultant bacteria were identified by Gram staining and biochemical testing.
After culture and identification of the different bacterial strains, sensitivity to commonly used antibiotics (including amoxicillin, amoxicillin–clavulanate, cefalotin and cefotaxime) was tested using the disc diffusion method, as recommended in 2008 by the Antibiogram Committee of the French Society of Microbiology. Pathogens with intermediate resistance were considered resistant to the relevant antibiotic. The sensitivity of S pneumoniae and H influenzae to amoxicillin and ceftriaxone was determined by assessing minimum inhibitory concentration, using the E-test. The Cefinase test (BIOMERIEUX, France SA) was used for H influenzae and M catarrhalis colonies to determine the capacity for beta-lactamase production. Strains of H influenzae were described as beta-lactamase-negative and ampicillin-resistant when beta-lactamase production was absent and the minimum amoxicillin inhibitory concentration was greater than 2 µg/ml.
Statistical analyses were performed using the Statistical Package for the Social Sciences software program (SPSS Inc, Chicago, Illinois, USA). The correlation between variables was evaluated using the chi-square test and Fisher's exact test. P values below 0.05 were considered to be statistically significant.
Results
A total of 62 children diagnosed with otitis media met the inclusion criteria. There were 40 boys (64.5 per cent) and 22 girls (35.5 per cent), with a mean ± standard deviation age of 4.21 ± 1.74 years. The majority were aged between three and five years (66 per cent). Around 61 per cent of the children had a history of smoking exposure, and 48.4 per cent had suffered recurrent upper respiratory tract infections. Amoxicillin–clavulanate was the commonest antibiotic used previously. Table I summarises patient characteristics.
Table I Patient characteristics
Data represent patient numbers (percentages) unless otherwise specified. *Per patient. †Per tested ear. SD = standard deviation; URTI = upper respiratory tract infection; prev = previous
One hundred and seven ears were tested for bacterial growth, resulting in positive bacterial cultures for 32 per cent (n = 34) of the tested ears. The most common cultured pathogen was H influenzae (n = 21, 62 per cent), followed by S pneumoniae (n = 9, 26 per cent) and moraxella species (n = 4, 12 per cent). Of those patients who had both ears tested, cultures were bilaterally positive in the majority.
Haemophilus influenzae had the greatest resistance profile to antibiotics commonly prescribed for acute otitis media; resistance findings were: amoxicillin, 81.0 per cent; amoxicillin–clavulanate, 52.4 per cent; cefalotin, 61.9 per cent; and cefotaxime, 19.0 per cent (p = 0.001) (Figure 1). The streptococcal resistance profile was low, with 11.11 per cent resistance to amoxicillin, to amoxicillin–clavulanate and to cefotaxime. Moraxella pathogens were sensitive to the latter two antibiotics, but showed 50 per cent resistance to amoxicillin. Fifty-two per cent of H influenzae strains were both beta-lactamase-negative and ampicillin-resistant, i.e. all strains had a negative Cefinase test, no evidence of beta-lactamase secretion, and a minimum amoxicillin inhibitory concentration of more than 2 µg/ml.
Fig. 1 Haemophilus influenzae sensititivity (sens) and resistance (res) profiles for beta-lactam antibiotics. Amox = amoxicillin; clav = clavulanate; cefalot = cefalotin; cefotax = cefotaxime
The mean duration of otitis media with effusion before surgery was four months. Assessment of the correlation between culture results and otitis media with effusion duration indicated that bacterial culture was more likely to be positive for otitis media with effusion cases of shorter duration (i.e. zero to two months (28.6 per cent culture positivity) and three to six months (20 per cent)), compared with those which had persisted for six months or more (0 per cent) (p < 0.001).
Fig. 2 Prevalence of culture-negative (neg) and culture-positive (pos) effusions, by patient age.
In addition, cultures were more likely to be positive in patients younger than four years of age, compared with older patients (culture positivity rates were 40 per cent for two-year-olds, 50 per cent for three-year-olds, 28.6 per cent for four-year-olds and 15.4 per cent for five-year-olds) (p = 0.015).
We found a high risk of developing H influenzae antibiotic resistance among children with a history of exposure to smoking (p = 0.001), recurrent upper respiratory tract infections (p = 0.001) and previous antibiotic treatment (p = 0.005). In H influenzae cases with exposure to smoking, 88.9, 55.6 and 66.7 per cent showed resistance to amoxicillin, amoxicillin–clavulanate and cefalotin, respectively. Similar findings were observed for children with a history of recurrent upper respiratory tract infection (Table II).
Table II Correlation between H influenzae antibiotic resistance and patient characteristics
amox = amoxicillin; clav = clavulanate; cefalot = cefalotin; cefotax = cefotaxime; exp = exposure; rec URTI = recurrent upper respiratory tract infections; prev Ab Rx = previous antibiotic therapy
Children with a mean age of 4 ± 1.5 years were at high risk of developing H influenzae resistance to amoxicillin, compared with those with a mean age of 2.5 ± 0.6 years (p = 0.006). No statistically significant correlation was observed between age and the development of resistance to other antibiotics.
Discussion
The bacterial types and prevalences identified in our cases of otitis media with effusion were similar to those reported in the world literature (i.e. 20–60 per cent).Reference Giebink, Juhn, Weber and Le2–Reference Pereira, Pereira, Cantarelli and Costa5 A higher incidence of positive culture was found to correlate with younger age (less than four years) and shorter duration of otitis media with effusion (six months or less). These findings are in keeping with the definition of otitis media with effusion as the persistence of fluid in the middle ear after resolution of acute otitis media, rather than eustachian tube dysfunction. Otitis media with effusion had previously been regarded as a strictly inflammatory process with a sterile effusion, until Senturia and colleagues' 1958 report of bacteria in otitis media with effusion redefined previously accepted concepts.Reference Senturia, Gessert, Carr and Baumann14
Our study findings are in agreement with other reports regarding the high prevalence, within otitis media with effusion fluid, of H influenzae followed by S pneumoniae and M catarrhalis, albeit with slightly varying proportions.Reference Pereira, Pereira, Cantarelli and Costa5–Reference El-Shamy11, Reference Mills, Uttley and McIntyre17–Reference Dagan, Abramson, Leibovitz, Lang, Goshen and Greenberg21 These minor differences from other studies may be attributed to varying inclusion criteria, sample sizes, microbiological methodology and geographical areas. Two similar Lebanese studies investigating the bacterial aetiology of paediatric otitis media with effusion have found that H influenzae was the most prevalent bacteria.Reference Hadi and Matar15, Reference Matar, Sidani, Fayad and Hadi16
Our patients' low incidence of S pneumoniae within otitis media effusion fluid may be due to the increased prevalence of vaccination against S pneumoniae strains. Furthermore, our patients' low incidence of moraxella strains was similar to incidences reported in other studies, and may be related to the high sensitivity and decreased resistance of this fragile organism to most commonly used antibiotics.
Several studies, in different countries, have shown that S pneumoniae and H influenzae are among the most frequent causative organisms for acute otitis media, and that bacterial resistance varies considerably over time and geographical region, as summarised in Table III.Reference Pereira, Pereira, Cantarelli and Costa5–Reference Li, Chiu, Hsu, Lee, Hwang and Huang7, Reference Kilpi, Herva, Kaijalainen, Syrjanem and Takala9 The streptococcal resistance profile to penicillin was low in our study; this could be related to the increased prevalence of vaccination against these bacterial strains in Lebanon. The H influenzae strains isolated from otitis media with effusion cases showed a high incidence of resistance to two of the most commonly prescribed antibiotics used in the treatment of paediatric upper respiratory tract infections (i.e. amoxicillin and amoxicillin–clavulanate).
Table III Reports of OME bacterial prevalence and antibiotic resistance
*M catarrhalis; †S aureus; ‡all bacteria. OME = otitis media with effusion; amox = amoxicillin; clav = clavulanate; cefalot = cefalotin; cefotax = cefotaxime; pen = penicillin; trimeth = trimethoprim; intermed = intermediate
This is the first study to report beta-lactamase-negative, ampicillin-resistant H influenzae strains in Lebanon; these strains constituted approximately half of the H influenzae strains cultured in our study. This finding could be due to the overuse of antibiotics in Lebanon, where antibiotics are available without a prescription. Hamamoto et al. reported in 2005 that long-term exposure to antibiotics may significantly influence the bacterial genome of S pneumoniae isolated from middle-ear effusions.Reference Hamamoto, Gotoh, Nakajo, Shimoya, Kayama and Hasegawa22
Many authors have noted that levels of antibiotic resistance are increasing markedly among bacterial pathogens which are common causes of paediatric acute otitis media, such as S pneumoniae and H influenzae.Reference Mills, Uttley and McIntyre17–Reference Rodriguez, Schwartz and Thorne19, Reference Hamamoto, Gotoh, Nakajo, Shimoya, Kayama and Hasegawa22–Reference Fluit, Florijn, Verhoef and Milatovic28
Haemophilus influenzae resistance to ampicillin emerged in the early 1970s, and levels of resistance have increased steadily since that time.Reference Hoban and Felmingham23 In H influenzae, resistance to ampicillin and other beta-lactam antibiotics is generally limited either to production of a beta-lactamase or, in the case of beta-lactamase-negative, ampicillin-resistant strains, to the presence of altered penicillin binding proteins with lowered affinity for beta-lactams.Reference James and Reeves24, Reference Tristram, Jacobs and Appelbaum25 A very small proportion of strains possess both mechanisms, and are referred to as beta-lactamase-positive, amoxicillin–clavulanate-resistant strains. Our results for beta-lactamase-negative, ampicillin-resistant H influenzae strains are similar to findings from Japan, where almost 40 per cent of H influenzae isolates have been reported to be beta-lactamase-negative and ampicillin-resistant.Reference Tristram, Jacobs and Appelbaum25–Reference Matic, Bozdogan, Jacobs, Ubukata and Appelbaum27 It is important to note that our definition of beta-lactamase-negative, ampicillin-resistant H influenzae was taken from another study, and that there is currently no consensus on this definition.Reference Matic, Bozdogan, Jacobs, Ubukata and Appelbaum27
Few studies have evaluated the antimicrobial susceptibility patterns of common bacterial pathogens found in cases of paediatric otitis media with effusion. Over the past two decades, findings from such studies have shown that levels of resistance to penicillin and/or ampicillin are much higher in H influenzae than S pneumoniae, as summarised in Table III.Reference Mills, Uttley and McIntyre17–Reference Rodriguez, Schwartz and Thorne19 In 2006, Rosenblüt et al. showed that S pneumoniae resistance to penicillin did not extend to amoxicillin; strains with intermediate and high levels of resistance to penicillin were more common (22 and 4.5 per cent, respectively) than strains with intermediate and high levels of resistance to amoxicillin (18 and 0.5 per cent, respectively).Reference Rosenblüt, Santolaya, Gonzalez, Borel and Cofré20 On the other hand, Dagan and colleagues' 1996 study found that 40 per cent of S pneumoniae isolates were intermediately penicillin-resistant, and that acute otitis media caused by such penicillin-resistant S pneumoniae strains responded inadequately to oral cephalosporins, mainly cefaclor.Reference Dagan, Abramson, Leibovitz, Lang, Goshen and Greenberg21
Eradication of S pneumoniae colonisation of the pharynx, using several different antibiotic classes including but not limited to macrolides, could spread azithromycin-nonsusceptible S pneumoniae.Reference Toltzis, Dul and Blumer29 A 1998 review paper suggested the introduction of new antibiotic treatments for acute otitis media, due to increased H influenzae and M catarrhalis resistance to commonly used antibiotics.Reference Hoppe and Johnson30
The range of pathogens responsible for acute otitis media has changed over recent decades. This situation has given rise to recent guidelines and studies highlighting the benefit of pneumococcal conjugate vaccines in reducing the overall disease burden of severe and recurrent acute otitis media, as well as the importance of appropriate antibiotic usage to avoid recurrence of infection and the development of bacterial antibiotic resistance.Reference Pelton6, Reference Klein31, Reference Coker, Chan, Newberry, Limbos, Suttorp and Shekelle32
A recent systematic review of acute otitis media has shown that the microbiology of acute otitis media has changed following the introduction of the PCV7 heptavalent pneumococcal conjugate vaccine. Prior to the introduction of this vaccine, the prevalence of S pneumoniae and H influenzae varied from 33 to 48 per cent and from 41 to 43 per cent, respectively. Following introduction of the vaccine, prevalence of S pneumonia decreased to 23–31 per cent while prevalence of H influenza increased to 56–57 per cent.Reference Coker, Chan, Newberry, Limbos, Suttorp and Shekelle32
• Beta-lactamase-negative, ampicillin-resistant Haemophilus influenzae strains are present in Lebanese children with otitis media with effusion (OME)
• This study found a high prevalence of H influenzae strains resistant to commonly used antibiotics
• Antibiotic resistance correlated with previous recurrent upper respiratory tract infections and smoking exposure
• Greater public awareness of these influences is needed, given the high local prevalence of smoking and probable antibiotics overuse
• It is important to constantly monitor OME bacteria antibiotic resistance, to enable appropriate guidance of patient treatment
The development of new diagnostic technology could help improve clinical and microbial diagnosis of acute otitis media.Reference Klein31 Myringotomy with aspiration and tympanostomy tube insertion can include the collection of material for microbiological monitoring of otitis media. This could help define local antibiotic resistance trends, and identify populations at higher risk of infections caused by resistant pathogens.
Our study findings indicate that pathogens such as H influenzae and M catarrhalis are clinically important in acute otitis media, and that therapy should be directed against these pathogens, in addition to S pneumoniae.
Conclusion
Our study established a microbiological profile of bacterial organisms associated with otitis media with effusion, and their antibiotic resistance patterns. There was a high prevalence of H influenzae strains resistant to commonly used antibiotics.
The presence of bacterial antibiotic resistance correlated with recurrent previous upper respiratory tract infections and smoking exposure, emphasising the need to improve public awareness in this regard.
In addition, we identified a high prevalence of beta-lactamase-negative, ampicillin-resistant H influenzae strains, a finding which may have important clinical implications.
Overall, we suggest that ongoing evaluation of the resistance patterns of bacteria implicated in otitis media with effusion would provide physicians with better information, facilitating appropriate treatment of such cases.
Acknowledgement
The authors would like to acknowledge Dr Issam Saliba for his critical review of this paper.
