Antibiotics for otitis media with effusion ('glue ear') in children

Review question

This review compared the effects of oral antibiotics against placebo, no treatment or other therapies in children with otitis media with effusion (OME) or 'glue ear'.

Background

Glue ear is one of the most common conditions of early childhood. Glue ear means that there is fluid in the middle ear space behind the eardrum. This may cause hearing difficulties that may in turn affect children's behaviour, language and progress at school. In approximately one in three children with glue ear, bacteria are identified in the middle ear fluid. Therefore, people have suggested that antibiotics may be beneficial in children with glue ear.

Study characteristics

This review included evidence available up to 14 April 2016. In total 25 studies (3663 children) were eligible for inclusion. Two studies did not report on any of the outcomes of interest, leaving 23 studies (3258 children). Overall, we assessed most studies as being at low to moderate risk of bias. In the 23 studies many different antibiotics were used and the children were of different ages and had suffered from glue ear for various lengths of time. They looked at the benefits at various time points after the treatment was given.

Key results

The most important outcomes that we measured were the difference in the proportion of children who no longer had glue ear two to three months after the treatment was started and adverse effects of antibiotics (diarrhoea, vomiting or skin rash).

We found moderate quality evidence (six trials including 484 children) that children treated with oral antibiotics are more likely to have glue ear resolved two to three months after the treatment was started than those allocated to control treatment. The number of children needed to treat for one beneficial outcome (NNTB) was five. However, there is evidence (albeit of low quality; five trials, 742 children) indicating that children treated with oral antibiotics are more likely to experience diarrhoea, vomiting or skin rash than those allocated to control treatment. The number of children needed to treat for one harmful outcome (NNTH) was 20.

In respect of the secondary outcome of having glue ear resolved at any time point, we found low to moderate quality evidence from five of our analyses where we combined data from studies (meta-analyses), which included between two and 14 studies, of a beneficial effect of antibiotics, with a NNTB ranging from three to seven. Time periods ranged from 10 to 14 days to six months.

In terms of other secondary outcomes, only two trials (849 children) reported on hearing levels at two to four weeks and found conflicting results. None of the trials reported data on speech, language and cognitive development or quality of life. Low quality evidence did not show that oral antibiotics were associated with fewer ventilation tube (grommet) insertions (two trials, 121 children) or in adverse consequences for the tympanic membrane (ear drum) (one trial, 103 children). Low quality evidence indicated that children treated with oral antibiotics were less likely to have acute otitis media (ear infection) episodes within four to eight weeks (five trials, 1086 children; NNTB 18) and within six months after treatment was started (two trials, 199 children; NNTB 5). It should however be noted that the beneficial effect of oral antibiotics on ear infection episodes within four to eight weeks was no longer significant when studies with high risk of bias were excluded.

Quality of the evidence

Moderate quality evidence is available that children with glue ear do benefit from oral antibiotics in terms of resolving glue ear at various time points and reducing acute otitis media episodes during follow-up compared with control treatment. Low quality evidence is available that children treated with oral antibiotics are more likely to experience diarrhoea, vomiting and skin rash than those receiving the control treatment. Currently only two trials have assessed the impact of oral antibiotics on hearing and these showed conflicting results (low quality evidence). Low quality evidence did not show that oral antibiotics were associated with fewer ventilation tube insertions or in adverse consequences for the tympanic membrane.

Authors' conclusions: 

This review presents evidence of both benefits and harms associated with the use of oral antibiotics to treat children up to 16 years with OME. Although evidence indicates that oral antibiotics are associated with an increased chance of complete resolution of OME at various time points, we also found evidence that these children are more likely to experience diarrhoea, vomiting or skin rash. The impact of antibiotics on short-term hearing is uncertain and low quality evidence did not show that oral antibiotics were associated with fewer ventilation tube insertions. Furthermore, we found no data on the impact of antibiotics on other important outcomes such as speech, language and cognitive development or quality of life.

Even in situations where clear and relevant benefits of oral antibiotics have been demonstrated, these must always be carefully balanced against adverse effects and the emergence of bacterial resistance. This has specifically been linked to the widespread use of antibiotics for common conditions such as otitis media.

Read the full abstract...
Background: 

Otitis media with effusion (OME) is characterised by an accumulation of fluid in the middle ear behind an intact tympanic membrane, without the symptoms or signs of acute infection. Since most cases of OME will resolve spontaneously, only children with persistent middle ear effusion and associated hearing loss potentially require treatment. Previous Cochrane reviews have focused on the effectiveness of ventilation tube insertion, adenoidectomy, nasal autoinflation, antihistamines, decongestants and corticosteroids in OME. This review, focusing on the effectiveness of antibiotics in children with OME, is an update of a Cochrane review published in 2012.

Objectives: 

To assess the benefits and harms of oral antibiotics in children up to 18 years with OME.

Search strategy: 

The Cochrane ENT Information Specialist searched the ENT Trials Register; Central Register of Controlled Trials (CENTRAL 2016, Issue 3); PubMed; Ovid EMBASE; CINAHL; Web of Science; ClinicalTrials.gov; ICTRP and additional sources for published and unpublished trials. The date of the search was 14 April 2016.

Selection criteria: 

Randomised controlled trials comparing oral antibiotics with placebo, no treatment or therapy of unproven effectiveness in children with OME.

Data collection and analysis: 

We used the standard methodological procedures expected by Cochrane.

Main results: 

Twenty-five trials (3663 children) were eligible for inclusion. Two trials did not report on any of the outcomes of interest, leaving 23 trials (3258 children) covering a range of antibiotics, participants, outcome measures and time points for evaluation. Overall, we assessed most studies as being at low to moderate risk of bias.

We found moderate quality evidence (six trials including 484 children) that children treated with oral antibiotics are more likely to have complete resolution at two to three months post-randomisation (primary outcome) than those allocated to the control treatment (risk ratio (RR) 2.00, 95% confidence interval (CI) 1.58 to 2.53; number needed to treat to benefit (NNTB) 5). However, there is evidence (albeit of low quality; five trials, 742 children) indicating that children treated with oral antibiotics are more likely to experience diarrhoea, vomiting or skin rash (primary outcome) than those allocated to control treatment (RR 2.15, 95% CI 1.29 to 3.60; number needed to treat to harm (NNTH) 20).

In respect of the secondary outcome of complete resolution at any time point, we found low to moderate quality evidence from five meta-analyses, including between two and 14 trials, of a beneficial effect of antibiotics, with a NNTB ranging from 3 to 7. Time periods ranged from 10 to 14 days to six months.

In terms of other secondary outcomes, only two trials (849 children) reported on hearing levels at two to four weeks and found conflicting results. None of the trials reported data on speech, language and cognitive development or quality of life. Low quality evidence did not show that oral antibiotics were associated with a decrease in the rate of ventilation tube insertion (two trials, 121 children) or in tympanic membrane sequelae (one trial, 103 children), while low quality evidence indicated that children treated with antibiotics were less likely to have acute otitis media episodes within four to eight weeks (five trials, 1086 children; NNTB 18) and within six months post-randomisation (two trials, 199 children; NNTB 5). It should, however, be noted that the beneficial effect of oral antibiotics on acute otitis media episodes within four to eight weeks was no longer significant when we excluded studies with high risk of bias.