A systematic review and meta-analysis of antimicrobial resistance in paediatric acute otitis media

International Journal of Pediatric Otorhinolaryngology 123 (2019) 102–109

Contents lists available at ScienceDirect

International Journal of Pediatric Otorhinolaryngology journal homepage: www.elsevier.com/locate/ijporl

A systematic review and meta-analysis of antimicrobial resistance in paediatric acute otitis media

T

Michael W. Mathera,b, Michael Drinnana, John D. Perryc, Steven Powellb, Janet A. Wilsonb,d, Jason Powella,b,∗ a

Institute of Cellular Medicine, Newcastle University, Framlington Place, Newcastle Upon Tyne, NE2 4HH, UK Department of Otolaryngology, Freeman Hospital, Freeman Road, Newcastle Upon Tyne, NE7 7DN, UK c Department of Microbiology, Freeman Hospital, Freeman Road, Newcastle Upon Tyne, NE7 7DN, UK d Institute of Health and Society, Newcastle University, Richardson Road, Newcastle Upon Tyne, NE2 4AX, UK b

A R T I C LE I N FO

A B S T R A C T

Keywords: Otitis media Antimicrobial resistance Bacteriology

Objective of review: Acute otitis media (AOM) is the largest cause of antimicrobial prescriptions amongst children in developed countries. Excessive and inappropriate prescribing is known to drive antimicrobial resistance, but less is known of antimicrobial resistance in AOM-associated bacteria. Type of review & search strategy: We conducted a systematic review and meta-analysis of bacterial prevalence and antimicrobial resistance in studies of paediatric AOM identified from Ovid Medline, Embase and the Cochrane library. Results: From 48 unique studies, 15,871 samples were included. Only 0.67 (CI 0.63–0.71) of all ear samples grew a bacterial pathogen. The most common bacterial causes of AOM in children were Streptococcus pneumoniae 0.30 (CI 0.27–0.32), Haemophilus influenza 0.23 (CI 0.20–0.26), and Moraxella catarrhalis 0.05 (CI 0.04–0.06). Resistance patterns varied amongst organisms and antimicrobial agents. The pooled proportion of bacterial culture-positive episodes of AOM that could be effectively treated with amoxicillin was 0.85 (CI 0.76–0.94), erythromycin was 0.64 (0.48–0.78) and amoxicillin-clavulanate was 0.95 (CI 0.85–0.98). Conclusion: We have demonstrated the bacteriology and antimicrobial resistance patterns of AOM. Of samples which grew bacteria, on average approximately 15% of isolates demonstrated resistance to amoxicillin; a typical first-line agent. Greater understanding of local bacteriology and resistance patterns is needed to enable improved antimicrobial stewardship.

1. Introduction Acute otitis media (AOM) is defined as the presence of inflammation in the middle ear associated with an effusion, and accompanied by the rapid onset of signs and symptoms of an ear infection [1]. It is the single largest cause of infections amongst children [2] and is the most common cause for antimicrobial prescriptions for children in economically developed countries [1]. Many national guidelines for AOM recommend either immediate or delayed antimicrobial prescribing (amoxicillin in most circumstances), or observation with close follow-up [3456]. These guidelines are, in part, based on work which has shown that after two to three days of watchful waiting approximately 80% of children will spontaneously recover [7]. Important exceptions to this include; children younger than 2 years, those with bilateral AOM, and those with AOM and otorrhoea,

∗

where antibiotics may be more beneficial [8]. Despite these evidencebased guidelines, large scale studies from North America [910], Europe [11], and the UK [12] have demonstrated excessive and inconsistent antimicrobial prescribing in paediatric AOM in general practice and secondary care. There is a clear association between antimicrobial prescribing and the development of antimicrobial resistance [13], and the World Health Organisation (WHO) calculated that in Europe alone infections due to drug-resistant bacteria cause in excess of 25,000 deaths and cost at least 1.5 billion euros each year in direct healthcare costs and lost productivity [14]. Furthermore, there is also evidence that specifically associates antimicrobial use to the development of antimicrobial resistance in AOM, and demonstrates that this increases the likelihood of treatment failure [15]. Recent studies have sought to identify the pathogens responsible for

Corresponding author. Institute of Cellular Medicine, Newcastle University, Framlington Place, Newcastle upon Tyne, NE2 4HH, UK. E-mail address: [email protected] (J. Powell).

https://doi.org/10.1016/j.ijporl.2019.04.041 Received 4 December 2018; Received in revised form 16 March 2019; Accepted 30 April 2019 Available online 06 May 2019 0165-5876/ © 2019 Elsevier B.V. All rights reserved.

International Journal of Pediatric Otorhinolaryngology 123 (2019) 102–109

M.W. Mather, et al.

susceptibility of three bacterial strains, to four common antimicrobials: penicillin; amoxicillin; amoxicillin-clavulanate; and erythromycin. For the relevant subset of species we also assessed beta-lactamase production as a surrogate for resistance to penicillin and amoxicillin. In a number of studies, only a subgroup of all positive cultures was assessed for antimicrobial susceptibility; the size of this subgroup was used as the denominator in the meta-analysis. Overall antimicrobial effectiveness: In almost all cases of AOM the decision on antimicrobial choice will be made without any prior knowledge of the organism involved. We therefore assessed the potential effectiveness of each common antimicrobial agent against an unknown pathogen. For each study, we estimated the overall susceptibility to a specific antimicrobial as reported in that study. This was estimated as the susceptibility of a given pathogen, weighted by the proportion of cases where that pathogen was present. We acknowledge that under this model, bias can be introduced where a particular antimicrobial agent is not used uniformly against all pathogens, or where multiple pathogens are present in the same sample.

paediatric acute otitis media [16], and older studies have also investigated the overall effectiveness of antibiotics in children with acute otitis media [8]. We present a comprehensive review and meta-analysis, of both the microbiology and antimicrobial resistance of AOM organisms to commonly used antimicrobial agents, with the aim of informing responsible antimicrobial stewardship. 2. Methods 2.1. Systematic review A comprehensive literature search was performed using Medline, Embase and the Cochrane library up to and including January 2017. A keyword search was undertaken using the search terms ‘otitis media’ AND each of the following search terms: ‘aetiology’, ‘otopathogens’, ‘pathogens’, ‘microbiology’, ‘bacteria’, ‘anti-bacterial agents’, and ‘antibiotic resistance’. Search results were limited to those that were in the English language, human-only studies, and published from 1980 onwards. Duplicated articles were removed and abstracts were screened for relevance to bacteriology and/or antimicrobial resistance in acute otitis media in children (less than 18 years old). Relevant articles were read in full to extract data for the metaanalysis. Inclusion criteria comprised all studies which provided original, non-duplicated quantitative data about the prevalence of bacteria with or without the number of susceptible, intermediate, and resistant strains of each species. Sampling methods included those obtained by tympanocentesis and otorrhoea from an acute tympanic membrane perforation, or combination of these sampling techniques. Notable exclusions comprised absence of original data, investigations of native flora, investigations specifically considering recurrent or ‘treatment-failure’ acute otitis media, studies with overlapping data (in which case we include the most recent report), studies exclusively sampling tympanostomy tube otorrhoea (except in mixed methods studies in which we only included series where tympanostomy tube otorrhoea comprised a small, < 20%, proportion of samples), nasopharyngeal swab data, or omission of fundamental demographic information (e.g. number of participants in study). No studies were excluded on the basis of their chosen method of culturing pathogens. Where AOM data was part of a larger study, only data relating to AOM was extracted. In drug trials, only the initial tympanocentesis data (i.e. before participants received the antimicrobial drug under investigation) was included.

3. Results The literature search yielded 7598 articles following the key-word search. Studies were limited to English language, human only studies, and studies from 1980-present day. Following deduplication this provided 4249 unique articles. Abstracts were screened for relevance to bacteriology or antimicrobial resistance in acute otitis media in children, which identified 204 articles. The full texts and bibliographies were read, and 48 articles [17–28] [29–64], had quantitative data that could be included in the analysis (Fig. 1, Table 1). Of these, 8 were retrospective studies, 27 were prospective studies, 4 were a combination of both prospective and retrospective analysis, and 9 were drug trials. In total these articles provided data for 15,871 unique episodes of AOM. 3.1. Microbiology Our best estimate demonstrated a positive culture in 0.67 (CI 0.63–0.71) of all samples. The most commonly isolated bacterial species was Streptococcus pneumoniae; a Gram-positive diplococcus wellknown to colonize upper respiratory tract mucosa [65], found in 0.30 (CI 0.27–0.32) of samples. The second most prevalent species was Haemophilus influenzae, a Gram-negative coccobacillus [66], which was isolated in 0.23 (CI 0.20–0.26) of samples. At 0.05 (CI 0.04–0.06) of samples Moraxella catarrhalis, a Gram-negative diplococcus, was a less frequent participant, but was the third most commonly isolated species. These are summarised in Fig. 2. The frequency of co-detection of multiple species for each sample was rarely reported in the original studies thereby precluding further analysis of this.

2.2. Meta-analysis All extracted data from the literature was included in the quantitative analysis. Data was tabulated in Microsoft Excel, then analysed using R (Vienna, Austria), version 3.3.1. Meta-analysis was performed using function metaprop and reported using function forest from the meta library. Pooled statistics were created from random-effects and fixed effects (inverse variance) meta-analyses. Heterogeneity was consistently extremely high (median I [2] > 90%), and therefore we discuss results on the basis of the random-effects meta-analysis. Each study is given a similar weight, whereas in the fixed-effects analysis all patients are weighted equally and large studies dominate the analysis. The random effect analysis is the more conservative approach, with wider confidence intervals. The methods are described below, with notes on the specific complicating factors: Positive culture: For each species separately and for all species together, we report the number of positive cultures as a proportion of all children assessed. Since some children are positive for more than one species, or are positive for a rare bacterial species not included in our data, the species totals are not necessarily additive to give the overall total. Antimicrobial susceptibility: We calculated the antimicrobial

3.2. Antimicrobial susceptibility 3.2.1. Streptococcus pneumoniae For Gram-positive bacteria, such as Streptococcus pneumoniae, most of the literature reported the proportion of samples resistant to penicillin. Testing for other specific drugs, even commonly used compounds such as amoxicillin, was much less frequent. It is important to note that whilst amoxicillin is a penicillin-based compound, some work suggests penicillin resistance is not fully predictive of resistance to related drugs, such as amoxicillin [40], which we also found in our results (Table 2). 3.2.2. Haemophilus influenzae & Moraxella catarrhalis For Gram-negative bacteria, such as Haemophilus influenzae and Moraxella catarrhalis, resistance was most often reported as the proportion of isolates capable of beta-lactamase production. Beta-lactamase is an enzyme produced by bacteria which hydrolyses the betalactam ring structure of commonly used penicillin-based drugs, 103

International Journal of Pediatric Otorhinolaryngology 123 (2019) 102–109

M.W. Mather, et al.

Fig. 1. PRIMSA flow diagram of literature search strategy.

Streptococcus pneumoniae and Haemophilus influenzae. Of particular concern, Moraxella catarrhalis, though less prevalent, demonstrated near universal production of beta-lactamase and resistance to amoxicillin. Streptococcus pneumoniae demonstrated high resistance to erythromycin; a common first-line agent in penicillin allergy. A promising finding was that, whilst studies were limited, susceptibility of all three organisms to amoxicillin-clavulanate, commonly used as a second-line agent, remains very high at > 90% of culture positive isolates. The strengths of this meta-analysis include its comprehensive nature. We included 48 studies and some 15,871 unique episodes of AOM. However, there are a number of considerations required when interpreting these finding. The International Organisation for Standardisation (ISO) has divided classification of Streptococcus pneumoniae resistance into susceptible, intermediate, and resistant categories [68]. ‘Susceptible’ bacteria are inhibited in vitro by a drug concentration associated with a high probability of therapeutic cure. ‘Intermediate’ varieties are associated with an uncertain clinical effect. ‘Resistant’ bacteria are associated with a high likelihood of therapeutic failure. Based on these definitions, we took all samples reported in the AOM literature as intermediate or resistant as being ‘non-susceptible’ isolates, as even intermediately susceptible bacteria in AOM have been shown to have impaired bacteriological clearance in response to antimicrobial drugs [69]. These ISO classifications are defined by points on a spectrum known as minimum inhibitory concentration (MIC) ‘breakpoints’. However, as there is geographical variation in these breakpoints comparison between studies from different locations is subject to variability. In terms of sampling, most (n = 30, 62.5%) studies sampled exclusively via tympanocentesis, which has been identified as the ‘gold standard’ for middle ear fluid culture [70]. Some (n = 3, 6.3%) studies also identified bacteria exclusively by sampling otorrhoea from the

including amoxicillin, thereby conferring a resistant phenotype [67] (Table 2). 3.2.3. Overall antimicrobial effectiveness in AOM We have demonstrated variable resistance against commonly used antimicrobial agents by different AOM-causing bacteria. In most cases antimicrobial prescription decisions will be made without knowledge of the causative organism. We therefore assessed the effectiveness of each antimicrobial agent against all positive bacterial cultures (Table 2). We also examined a pooled estimate of non-susceptibility to penicillin over time and found no statistically significant trend (Fig. 3). 3.2.4. Factors affecting antimicrobial effectiveness in AOM We were unable to identify any obvious trends towards changes in antimicrobial resistance or bacteriology of AOM over the period covered in this meta-analysis. Subgroup analysis of study cohort age, sampling method and geographical location was severely limited by the data available. 4. Discussion Even when a bacteriological cause for AOM is confirmed, many first-line antimicrobial treatments for AOM demonstrate drug-resistance. Whilst the analysed data indicates a pathogenic bacterial species was isolated in two out of every three cases of AOM, it is possible that the children in the studies analysed are at the more severe end of the AOM spectrum as they have engaged with medical services and had an intervention, such as tympanocentesis. Of the positive cultures the bacteriology of AOM was dominated by Streptococcus pneumoniae and Haemophilus influenzae. Non-susceptibility to commonly used first-line agents was found to be high in both 104

International Journal of Pediatric Otorhinolaryngology 123 (2019) 102–109

M.W. Mather, et al.

Table 1 Descriptive statistics of studies included in quantitative analysis (ND=Not documented). Reference

Country of origin

Percentage (%) tympanocentesis sampling

Percentage (%) otorrhoea sampling

Percentage (%) tympanostomy tube otorrhoea sampling

Year study commenced

Year study ended

Lower age limit of participants (months)

Upper age limit of participants (months)

Number of samples in study

Douglas 1980 Karma 1987 Trujillo 1989 Arguedas 1991 Rodriguez 1995 Aronovitz 1996 Gehanno 1996 Arguedas 1998 Jacobs 1998 Heikkinen 1999 Block 2000 Dagan 2000 Commisso 2000 Dagan 2001 Gehanno 2001 Kilpi 2001 Leiberman 2001 Rosenblut 2001 Sih 2001 Broides 2002 Turner 2002 Arguedas 2003 Piglansky 2003 Oguz 2003 Hoberman 2005 Guven 2006 Rosenblut 2006 Sakran 2006 Patel 2007 Aguilar 2009 Brook 2009 Yano 2009 Grubb 2010 Parra 2011 Sierra 2011 Grevers 2012 Naranjo 2012 Casey 2013 Chen 2013 Falup-Pecurariu 2013 Al-Mazrou 2014 Intakorn 2014 Kung 2014 Abdelnour 2015 Ding 2015 Tamir 2015 Sillanpaa 2016 Yatsyshina 2016

Australia Finland Colombia USA USA USA France Costa Rica International USA USA International Argentina International France Finland Israel Chile Brazil Israel Israel Costa Rica Israel Turkey International Turkey Chile Israel USA Costa Rica USA Japan USA Mexico Colombia Germany Venezuela USA Taiwan Romania

100.0 100 100 100 97.5 100 100.0 100 80 100 100 100 100 100 86.6 100 100 100 100 100 100 81 100 100 100 98.1 100 100 100 100 – 100 100 82 85 24 90.1 100 – 67.0

0.0 – – – 2.5 – 0.0 – 18 – – – – – 13.4 – – – – – – 19 – – – 1.9 – – – – 100 – – 18 15 76 9.9 – 100 33.0

– – – – – – – – 2 – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

1977 1980 1979 ND 1994 1996 1992 1992 1994 1989 1996 1997 1996 1999 1987 1994 1998 1998 1990 1998 1995 1999 1999 1998 2001 2002 1998 2002 1989 2002 1998 2002 2005 2008 2008 2008 2008 2008 2011 2009

1978 1985 1985 ND 1995 1996 1993 1997 1995 1993 1996 1998 1997 1999 1997 1995 1999 1999 1995 2000 1999 2001 2001 2000 2002 2004 2002 2004 1998 2007 2006 2004 2009 2009 2009 2010 2009 2010 2012 2011

4 0 0.1 ND 12 24 0 4 3 2 6 3.9 1 3 3 2 3 3 2 3 0 4 3 6 6 6 3 0 2 2 5 0 ND 3 3 3 0 6 0 0

168 2 132 ND 72 180 71 144 60 84 144 106.6 24 48 36 24 32 ND 60 36 2 123 36 144 30 144 108 2 84 92 144 120 ND 60 60 60 60 36 215 59

103 155 111 122 159 169 303 398 915 815 177 238 367 521 2149 772 73 170 300 145 137 102 78 78 730 180 543 68 982 880 124 1092 184 121 99 100 91 208 69 212

Saudi Arabia Thailand Taiwan Costa Rica China Israel Finland Russia

69 91 100 90.1 – 76 78 100

31.0 9 – 9.9 100 24 5 –

– – – – – – 17 –

2009 2008 2009 2010 2011 2008 2010 2011

2011 2009 2011 2012 2013 2013 2011 2013

3 5 4 3 0 0 6 0

60 59 213 59 216 72 39 59

75 118 151 456 229 295 91 216

to have higher rates of bacterial detection compared to the average over all studies, which suggests that culture-dependent detection might underestimate the prevalence of positive bacterial infection. Whilst changing from culture-based methods to PCR may alter the detection of bacteria, this technique will not alter the ways in which antimicrobial resistance is assessed. There is also inconsistency between included studies in terms of diagnostic criteria for AOM. Other studies have found that diagnoses of AOM are not always supported by positive physical examination findings [74]; and therefore it is plausible that a number of studies which did not demonstrate a positive bacterial culture result were not in fact AOM. The possibility of this is an inherent weakness of a retrospective study on the subject. Consideration also ought to be given to the lack of quantification of variables known to alter antimicrobial resistance, such as previous antimicrobial therapy [13]. Indeed, a substantial problem affecting all studies of antimicrobial resistance is that whilst a prescription may have been provided to a patient; few studies confirm that the antibiotics

external auditory canal following acute tympanic perforation; which potentially risks contamination with the native flora of the external auditory canal [71]. The remaining 15 studies (31.3%) involved mixed sampling methods. Studies exclusively examining tympanostomy tube otorrhoea were excluded on the basis that they reflect a chronically perforated tympanic membrane and will likely be contaminated with migration of commensal species from the external auditory canal [72]. Data from nasopharyngeal swabs were also excluded as they are felt to be unrepresentative of the middle ear [70]: indeed a recent systematic review found a wide range of concordance between nasopharyngeal samples and middle ear samples - from 68 up to 97% [73]. Studies specifically investigating recurrent or ‘treatment failure’ AOM were excluded on the basis that they introduce a confounding variable to the analysis and would likely warrant specific investigation in a separate study. Whilst most studies to date utilised culture-based methods of bacterial identification, one more recent study (Yatyshina et al., 2016) used polymerase chain reaction (PCR) technology [63]. This study appeared 105

International Journal of Pediatric Otorhinolaryngology 123 (2019) 102–109

M.W. Mather, et al.

Fig. 2. Bacterial prevalence Forrest plots.

106

International Journal of Pediatric Otorhinolaryngology 123 (2019) 102–109

M.W. Mather, et al.

Table 2 Measures of antibiotic effectiveness for commonly tested antibiotic agents against the most common acute otitis media pathogens. Each cell shows: (top) estimate of effectiveness; (centre) 95% confidence intervals; (bottom) number of studies (st) and patients (px) pooled.

Streptococcus pneumoniae

Haemophilus influenza

Moraxella catarrhalis

All samples (excluding negative cultures)

a

Amoxicillin

Penicillin

Amoxicillin-clavulanate

Erythromycin

0.86 [0.69 to 0.95] 9 st, 1121 px 0.82 [0.67 to 0.92] 5 st, 238 px 0.12 [0.01 to 0.65] 3 st, 58 px 0.85 [0.73 to 0.93] 11 st, 1417 px

0.60 [0.51 to 0.68] 29 st, 2704 px 0.43 [0.05 to 0.92] 3 st, 46 px 0.13 [0.01 to 0.63] 3 st, 63 px 0.56 [0.47 to 0.65] 29 st, 2813 px

0.93 [0.81 to 0.98] 8 st, 888 px 0.98 [0.88 to 1.00] 7 st, 564 px 0.98 [0.91 to 1.00] 3 st, 77 px 0.95 [0.85 to 0.98] 10 st, 1529 px

0.64 [0.48 to 0.78] 14 st, 976 px 0.53 [0.30 to 0.74] 1 st, 17 pxa 0 st, 0 px

Beta lactamase non-producer

0.71 [0.61 to 0.79] 25 st, 2280 px 0.07 [0.03 to 0.17] 13 st, 205 px

0.64 [0.48 to 0.78] 14 st, 993 px

Data based on the single study entered into the meta-analysis.

bacteria compared to our pooled data, however we were unable to control for these factors. There is also the challenge of consistent patient selection across so many studies. Although a definition of AOM is provided [1], not all papers had identical inclusion criteria. Two large trials (Hoberman NEJM 2016 & Tahtinen NEJM 2011) suggested that use of very stringent inclusion criteria revealed a high percentage of pathogen isolated compared to the mean values reported in this analysis. Whilst we were unable to identify a statistically significant rise in non-susceptibility to penicillin over time amongst all species tested (Fig. 3) factors such as disparate geography and variable vaccination schedules may reduce the reliability of this for any one specific location and time point. The introduction of different guidelines over time may also affect the estimate. We include children and adolescents up to 18 years because many of the published studies presented this range as their inclusion criteria. Furthermore, patient age was often not defined beyond the descriptor of less than 18 years so further data extraction and subgroup analysis was not possible. AOM is most commonly a disease of early childhood [1]. As anatomy and immunology changes during childhood, it is possible that these factors alter AOM bacteriology and antimicrobial resistance

were actually dispensed and administered. As such, the stated rates of antimicrobial prescribing may be overstated compared to the true figures. Further confounding factors include the different geographical locations of the studies; each with differing rates of prescribing and unique epidemiology of bacteria [75]. We attempted to perform subgroup analysis of antimicrobial resistance by continent but, whilst effects were suggested, the small group sizes and potential confounding factors preclude drawing reliable conclusions. Similarly, the heterogeneity of guidelines for the management of AOM in different locations at different points in time precluded further analysis of this, but would be valuable topics for future research. Another consideration is the variable rates of, and often lack of documentation of, pneumococcal (PCV) and Haemophilus influenzae type B (Hib) vaccination and the impact on resistance. Although we were unable to control for this variable in countries with universal vaccination against Haemophilus influenzae it may be that other species are relatively more prevalent causes of AOM, which will have correspondingly different levels of antimicrobial resistance. Similarly, countries in which antimicrobial drugs, such as amoxicillin, are available without prescription, one may find higher levels of drug-resistant

Fig. 3. Time course of non-susceptibility to penicillin amongst all bacterial isolates. 107

International Journal of Pediatric Otorhinolaryngology 123 (2019) 102–109

M.W. Mather, et al.

Competing financial interests disclosure

patterns as a child ages, however such data was not available for analysis. Other important considerations are the paucity of data on biofilms in the included studies, which have known importance in OM [76]. One must also consider the selection of only English language studies which may have excluded relevant articles in other languages.

The authors have no financial relationships relevant to this article to disclose.

4.1. Implications of the findings

The authors have no conflicts of interest relevant to this article to disclose.

Conflicts of interest

It was been well-described that acute otitis media may be associated with diagnostic uncertainty. Emphasis must therefore be made on treating true episodes of AOM with an effective antimicrobial agent for an appropriate duration. However, this data suggests many prescriptions adhering to present guidelines are likely to be microbiologically ineffective and therefore of no clinical benefit. For example, our global estimates suggest that Erythromycin is frequently ineffective – and yet still confers risks of adverse drug reactions. This therefore suggests its role as a first line agent may be limited. A meta-analysis by Rosenfeld et al. demonstrated that over 60% of cases of AOM resolve spontaneously after 24 h, and 80% within 2–3 days, when not treated with antimicrobial drugs [7]. Some clinicians might consider prescribing them in an attempt to mitigate the risk of developing serious complications of AOM. However, a retrospective analysis of the records of 2.5 million children found that while administration of antimicrobial therapy in primary care reduced the incidence of mastoiditis, 4831 episodes of otitis media needed to be treated with antimicrobial drugs to prevent 1 case of mastoiditis [77]. Ineffective treatment of resistant bacteria confers all the risks of drug side-effects for the patient without providing any benefits. These risks are not inconsiderable; a recent Cochrane review found that one in fourteen patients will experience an adverse event, including vomiting, diarrhoea, or a rash [78]. Furthermore, continued use of these agents may continue to drive further selection of resistant bacteria.

Acknowledgements Not applicable. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ijporl.2019.04.041. References [1] M.M. Rovers, A.G.M. Schilder, G.A. Zielhuis, R.M. Rosenfeld, Otitis media, Lancet 363 (2004) 465–473. [2] V.M. Freid, D.M. Makuc, R.N. Rooks, Ambulatory health care visits by children: principal diagnosis and place of visit, Vital Health Stat 13 (1998) 1–23. [3] A.S. Lieberthal, et al., The diagnosis and management of acute otitis media, Pediatrics 131 (2013) e964–e999. [4] NICE, Respiratory Tract Infections – Antibiotic Prescribing, (2008) London. [5] N.R. Le Saux, J. . Management of acute otitis media in children six months of age and older, Paediatr. Child Health 21 (2016) 39–44. [6] R. Azria, B. Barry, E. Bingen, J. Cavallo, C. Chidiac, M. Francois, E.P.J. Grimprel, E. Varon, A. Wollner, R. Cohen, Systemic antibiotherapy in routine practice for upper respiratory tract infections in adults and children, Med. Maladies Infect. 42 (2012) 460–487. [7] R.M. Rosenfeld, D. Kay, Natural history of untreated otitis media, Laryngoscope 113 (2003) 1645–1657. [8] M.M. Rovers, et al., Antibiotics for acute otitis media: a meta-analysis with individual patient data, Lancet 368 (2006) 1429–1435. [9] M.E. Pichichero, Preferred antibiotics for treatment of acute otitis media: comparison of practicing pediatricians, general practitioners, and otolaryngologists, Clin. Pediatr. 44 (2005) 575–578. [10] L.N. McEwen, R. Farjo, B. Foxman, Antibiotic prescribing for otitis media: how well does it match published guidelines? Pharmacoepidemiol. Drug Saf. 12 (2003) 213–219. [11] A.I.O. Plasschaert, M.M. Rovers, A.G.M. Schilder, T.J.M. Verheij, E. Hak, Trends in doctor consultations, antibiotic prescription, and specialist referrals for otitis media in children: 1995-2003, Pediatrics 117 (2006) 1879–1886. [12] J.I. Hawker, et al., Trends in antibiotic prescribing in primary care for clinical syndromes subject to national recommendations to reduce antibiotic resistance, UK 1995-2011: analysis of a large database of primary care consultations, J. Antimicrob. Chemother. 69 (2014) 3423–3430. [13] H. Goossens, M. Ferech, R. Vander Stichele, M. Elseviers, E.P. Group, Outpatient antibiotic use in Europe and association with resistance: a cross-national database study, Lancet 365 (2005) 579–587. [14] W.H. Organisation, Global Action Plan on Antimicrobial Resistance, World Health Organisation, 2015. [15] I. Brook, A.E. Gober, Resistance to antimicrobials used for therapy of otitis media and sinusitis: effect of previous antimicrobial therapy and smoking, Ann. Otol. Rhinol. Laryngol. 108 (1999) 645–647. [16] C.C. Ngo, H.M. Massa, R.B. Thornton, A.W. Cripps, Predominant bacteria detected from the middle ear fluid of children experiencing otitis media: a systematic review, PLoS One 11 (2016) e0150949. [17] M.R. Jacobs, R. Dagan, P.C. Appelbaum, D.J. Burch, Prevalence of antimicrobialresistant pathogens in middle ear fluid: multinational study of 917 children with acute otitis media, Antimicrob. Agents Chemother. 42 (1998) 589–595. [18] L. Aguilar, et al., Microbiology of the middle ear fluid in Costa Rican children between 2002 and 2007, Int. J. Pediatr. Otorhinolaryngol. 73 (2009) 1407–1411. [19] Y. Ding, et al., Etiology and epidemiology of children with acute otitis media and spontaneous otorrhea in Suzhou, China, Pediatr. Infect. Dis. J. 34 (2015) e102–e106. [20] P. Gehanno, et al., Microbiology of otitis media in the Paris, France, area from 1987 to 1997, Pediatr. Infect. Dis. J. 20 (2001) 570–573. [21] P. Intakorn, et al., Haemophilus influenzae type b as an important cause of culturepositive acute otitis media in young children in Thailand: a tympanocentesis-based, multi-center, cross-sectional study, BMC Pediatr. 14 (2014) 157. [22] L. Naranjo, et al., Non-capsulated and capsulated Haemophilus influenzae in children with acute otitis media in Venezuela: a prospective epidemiological study, BMC Infect. Dis. 12 (2012) 40. [23] A. Sierra, et al., Non-typeable Haemophilus influenzae and Streptococcus

5. Conclusions Commonly used first-line antimicrobial agents are unlikely to confer any positive effect in many cases of paediatric AOM. Firstly, due to the frequently non-bacterial nature of the condition; and secondly, the evidence of bacterial resistance to commonly used first-line antimicrobial agents. Author contributions MM was responsible for the conception and content of the article, the database searches, data interpretation, and preparation of the manuscript. MD performed the statistical analysis, data interpretation, and assisted in manuscript preparation. JDP assisted with microbiological data interpretation and manuscript preparation. SP contributed clinical interpretation of data and manuscript preparation. JW contributed clinical interpretation of data and manuscript preparation. JP was responsible for the conception and content of the article, data interpretation, and preparation of the manuscript. All authors approved the final manuscript. Data availability The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. Funding source This work received no specific funding. 108

International Journal of Pediatric Otorhinolaryngology 123 (2019) 102–109

M.W. Mather, et al.

[24]

[25]

[26]

[27]

[28]

[29]

[30] [31]

[32]

[33]

[34]

[35] [36]

[37] [38]

[39] [40]

[41] [42]

[43]

[44]

[45] [46]

[47]

[48] [49]

[50]

Immunol. Infect. 46 (2013) 382–388. [51] R. Dagan, et al., Bacteriologic and clinical efficacy of amoxicillin/clavulanate vs. azithromycin in acute otitis media, Pediatr. Infect. Dis. J. 19 (2000) 95–104. [52] R.M. Douglas, H. Miles, D. Hansman, B. Moore, D.T. English, Microbiology of acute otitis media with particular reference to the feasibility of pneumococcal immunization, Med. J. Aust. 1 (1980) 263–266. [53] O. Falup-Pecurariu, et al., Pneumococcal acute otitis media in infants and children in central Romania, 2009-2011: microbiological characteristics and potential coverage by pneumococcal conjugate vaccines, Int. J. Infect. Dis. 17 (2013) e702–e706. [54] P. Gehanno, et al., Evaluation of nasopharyngeal cultures for bacteriologic assessment of acute otitis media in children, Pediatr. Infect. Dis. J. 15 (1996) 329–332. [55] M.S. Grubb, D.C. Spaugh, Microbiology of acute otitis media, puget sound region, 2005-2009, Clin. Pediatr. 49 (2010) 727–730. [56] A. Hoberman, et al., Large dosage amoxicillin/clavulanate, compared with azithromycin, for the treatment of bacterial acute otitis media in children, Pediatr. Infect. Dis. J. 24 (2005) 525–532. [57] P.H. Karma, J.S. Pukander, M.M. Sipila, T.H. Vesikari, P.W. Gronroos, Middle ear fluid bacteriology of acute otitis media in neonates and very young infants, Int. J. Pediatr. Otorhinolaryngol. 14 (1987) 141–150. [58] Y.H. Kung, et al., Bacterial etiology of acute otitis media in the era prior to universal pneumococcal vaccination in Taiwanese children, J. Microbiol. Immunol. Infect. 47 (2014) 239–244. [59] A. Leiberman, et al., Bacteriologic and clinical efficacy of trimethoprim-sulfamethoxazole for treatment of acute otitis media, Pediatr. Infect. Dis. J. 20 (2001) 260–264. [60] W. Sakran, et al., Acute otitis media in infants less than three months of age: clinical presentation, etiology and concomitant diseases, Int. J. Pediatr. Otorhinolaryngol. 70 (2006) 613–617. [61] T.M. Sih, Acute otitis media in Brazilian children: analysis of microbiology and antimicrobial susceptibility, Ann. Otol. Rhinol. Laryngol. 110 (2001) 662–666. [62] S. Sillanpaa, M. Sipila, H. Hyoty, M. Rautiainen, J. Laranne, Antibiotic resistance in pathogens causing acute otitis media in Finnish children, Int. J. Pediatr. Otorhinolaryngol. 85 (2016) 91–94. [63] S. Yatsyshina, et al., Detection of respiratory pathogens in pediatric acute otitis media by PCR and comparison of findings in the middle ear and nasopharynx, Diagn. Microbiol. Infect. Dis. 85 (2016) 125–130. [64] H. Trujillo, R. Callejas, G.I. Mejia, L. Castrillon, Bacteriology of middle ear fluid specimens obtained by tympanocentesis from 111 Colombian children with acute otitis media, Pediatr. Infect. Dis. J. 8 (1989) 361–363. [65] A. Kadioglu, J.N. Weiser, J.C. Paton, P.W. Andrew, The role of Streptococcus pneumoniae virulence factors in host respiratory colonization and disease, Nat. Rev. Microbiol. 6 (2008) 288–301. [66] J. Van Eldere, M.P.E. Slack, S. Ladhani, A.W. Cripps, Non-typeable Haemophilus influenzae, an under-recognised pathogen, Lancet Infect. Dis. 14 (2014) 1281–1292. [67] S.M. Drawz, R.A. Bonomo, Three decades of beta-lactamase inhibitors, Clin. Microbiol. Rev. 23 (2010) 160–201. [68] ISO - "Clinical Laboratory Testing and in Vitro Diagnostic Test Systems — Susceptibility Testing of Infectious Agents and Evaluation of Performance of Antimicrobial Susceptibility Test Devices". [69] R. Dagan, et al., Impaired bacteriologic response to oral cephalosporins in acute otitis media caused by pneumococci with intermediate resistance to penicillin, Pediatr. Infect. Dis. J. 15 (1996) 980–985. [70] R. Kaur, K. Czup, J.R. Casey, M.E. Pichichero, Correlation of nasopharyngeal cultures prior to and at onset of acute otitis media with middle ear fluid cultures, BMC Infect. Dis. 14 (2014). [71] N. Principi, P. Marchisio, C. Rosazza, C.S. Sciarrabba, S. Esposito, Acute otitis media with spontaneous tympanic membrane perforation, Eur. J. Clin. Microbiol. Infect. Dis. 36 (2017) 11–18. [72] J. Dohar, Microbiology of otorrhea in children with tympanostomy tubes: implications for therapy, Int. J. Pediatr. Otorhinolaryngol. 67 (2003) 1317–1323. [73] T.M. van Dongen, et al., Evaluation of concordance between the microorganisms detected in the nasopharynx and middle ear of children with otitis media, Pediatr. Infect. Dis. J. 32 (2013) 549–552. [74] D.L. Brinker, W.L. MacGeorge, Hackman, Diagnostic accuracy, prescription behavior, and watchful waiting efficacy for pediatric acute otitis media, Clin. Pediatr. (N. Y.) 58 (1) (2019) 60–65. [75] T.P. Van Boeckel, et al., Global antibiotic consumption 2000 to 2010: an analysis of national pharmaceutical sales data, Lancet Infect. Dis. 14 (2014) 742–750. [76] Q. Vermee, R. Cohen, C. Hays, E. Varon, S. Bonacorsi, S. Bechet, F. Thollot, F. Corrard, C. Poyart, C. Levy, J. Raymond, Biofilm production by Haemophilus influenzae and Streptococcus pneumoniae isolated from the nasopharynx of children with acute otitis media, BMC Infect. Dis. 11 (1) (2019) 44 19. [77] P.L. Thompson, et al., Effect of antibiotics for otitis media on mastoiditis in children: a retrospective cohort study using the United Kingdom general practice research database, Pediatrics 123 (2009) 424–430. [78] R.P. Venekamp, S.L. Sanders, P.P. Glasziou, C.B. Del Mar, M.M. Rovers, Antibiotics for acute otitis media in children, Cochrane Database Syst. Rev. (6) (2015) CD000219, , https://doi.org/10.1002/14651858.CD000219.pub4.

pneumoniae as primary causes of acute otitis media in colombian children: a prospective study, BMC Infect. Dis. 11 (2011). D. Turner, et al., Acute otitis media in infants younger than two months of age: microbiology, clinical presentation and therapeutic approach, Pediatr. Infect. Dis. J. 21 (2002) 669–674. R. Commisso, F. Romero-Orellano, P.B. Montanaro, F. Romero-Moroni, R. RomeroDiaz, Acute otitis media: bacteriology and bacterial resistance in 205 pediatric patients, Int. J. Pediatr. Otorhinolaryngol. 56 (2000) 23–31. T. Kilpi, E. Herva, T. Kaijalainen, R. Syrjanen, A.K. Takala, Bacteriology of acute otitis media in a cohort of Finnish children followed for the first two years of life, Pediatr. Infect. Dis. J. 20 (2001) 654–662. M. Guven, et al., Bacterial etiology of acute otitis media and clinical efficacy of amoxicillin-clavulanate versus azithromycin, Int. J. Pediatr. Otorhinolaryngol. 70 (2006) 915–923. G. Grevers, et al., Identification and characterization of the bacterial etiology of clinically problematic acute otitis media after tympanocentesis or spontaneous otorrhea in German children, BMC Infect. Dis. 12 (2012). A. Abdelnour, et al., Etiology and antimicrobial susceptibility of middle ear fluid pathogens in Costa Rican children with otitis media before and after the introduction of the 7-valent pneumococcal conjugate vaccine in the national immunization program: acute otitis media microbiology in Costa Rican children, Medicine 94 (2015) e320. A. Arguedas, et al., Microbiology of acute otitis media in Costa Rican children, Pediatr. Infect. Dis. J. 17 (1998) 680–684. A.G. Arguedas, M. Zaleska, H.R. Stutman, J.L. Blumer, C.S. Hains, Comparative trial of cefprozil vs. amoxicillin clavulanate potassium in the treatment of children with acute otitis media with effusion, Pediatr. Infect. Dis. J. 10 (1991) 375–380. J.R. Casey, R. Kaur, V.C. Friedel, M.E. Pichichero, Acute otitis media otopathogens during 2008 to 2010 in Rochester, New York, Pediatr. Infect. Dis. J. 32 (2013) 805–809. R. Dagan, et al., Bacteriologic and clinical efficacy of high dose amoxicillin/clavulanate in children with acute otitis media, Pediatr. Infect. Dis. J. 20 (2001) 829–837. F. Oguz, et al., Etiology of acute otitis media in childhood and evaluation of two different protocols of antibiotic therapy: 10 days cefaclor vs. 3 days azitromycin, Int. J. Pediatr. Otorhinolaryngol. 67 (2003) 43–51. L. Piglansky, et al., Bacteriologic and clinical efficacy of high dose amoxicillin for therapy of acute otitis media in children, Pediatr. Infect. Dis. J. 22 (2003) 405–412. W.J. Rodriguez, R.H. Schwartz, M.M. Thorne, Increasing incidence of penicillinand ampicillin-resistant middle ear pathogens, Pediatr. Infect. Dis. J. 14 (1995) 1075–1078. A. Rosenblut, et al., Bacterial and viral etiology of acute otitis media in Chilean children, Pediatr. Infect. Dis. J. 20 (2001) 501–507. S.O. Tamir, et al., Changing trends of acute otitis media bacteriology in central Israel in the pneumococcal conjugate vaccines era, Pediatr. Infect. Dis. J. 34 (2015) 195–199. M.M. Parra, et al., Bacterial etiology and serotypes of acute otitis media in Mexican children, Vaccine 29 (2011) 5544–5549. A. Rosenblut, M.E. Santolaya, P. Gonzalez, C. Borel, J. Cofre, Penicillin resistance is not extrapolable to amoxicillin resistance in Streptococcus pneumoniae isolated from middle ear fluid in children with acute otitis media, Ann. Otol. Rhinol. Laryngol. 115 (2006) 186–190. T. Heikkinen, M. Thint, T. Chonmaitree, Prevalence of various respiratory viruses in the middle ear during acute otitis media, N. Engl. J. Med. 340 (1999) 260–264. J.A. Patel, D.T. Nguyen, K. Revai, T. Chonmaitree, Role of respiratory syncytial virus in acute otitis media: implications for vaccine development, Vaccine 25 (2007) 1683–1689. H. Yano, et al., Detection of respiratory viruses in nasopharyngeal secretions and middle ear fluid from children with acute otitis media, Acta Otolaryngol. 129 (2009) 19–24. K.A. Al-Mazrou, A.M. Shibl, W. Kandeil, J.Y. Pircon, C. Marano, A prospective, observational, epidemiological evaluation of the aetiology and antimicrobial susceptibility of acute otitis media in Saudi children younger than 5years of age, Journal of Epidemiology and Global Health 4 (2014) 231–238. A. Arguedas, et al., Microbiology of otitis media in Costa Rican children, 1999 through 2001, Pediatr. Infect. Dis. J. 22 (2003) 1063–1068. G. Aronovitz, A multicenter, open label trial of azithromycin vs. amoxicillin/clavulanate for the management of acute otitis media in children, Pediatr. Infect. Dis. J. 15 (1996) S15–S19. S.L. Block, J.A. Hedrick, J. Kratzer, M.A. Nemeth, K.J. Tack, Five-day twice daily cefdinir therapy for acute otitis media: microbiologic and clinical efficacy, Pediatr. Infect. Dis. J. 19 (2000) S153–S158. A. Broides, et al., Cytology of middle ear fluid during acute otitis media, Pediatr. Infect. Dis. J. 21 (2002) 57–61. I. Brook, A.E. Gober, Bacteriology of spontaneously draining acute otitis media in children before and after the introduction of pneumococcal vaccination, Pediatr. Infect. Dis. J. 28 (2009) 640–642. Y.J. Chen, Y.C. Hsieh, Y.C. Huang, C.H. Chiu, Clinical manifestations and microbiology of acute otitis media with spontaneous otorrhea in children, J. Microbiol.

109