Summary of the evidence

The recommendations in this guideline are based on the evidence identified, which was for children and young people under 18 years. [2018]

Non‑antimicrobial treatments

Oral analgesia (paracetamol and ibuprofen)

  • Paracetamol and ibuprofen were both more effective than placebo in reducing pain at 48 hours in children with acute otitis media (number needed to treat [NNT] 6 to 7 [range 4 to 27] for no pain at 48 hours; low to moderate quality evidence). This was based on a systematic review of randomised controlled trials (RCTs; Sjoukes et al. 2016). There were no significant differences in fever at 48 hours with paracetamol or ibuprofen compared with placebo (very low quality evidence). [2018]

  • No significant differences were found between paracetamol and ibuprofen for pain or fever at various time points (very low to low quality evidence). Furthermore, using ibuprofen and paracetamol in combination was no more effective than paracetamol alone, although this was based on very small numbers of children (very low to low quality evidence; Sjoukes et al. 2016). [2018]

  • Adverse events for paracetamol and ibuprofen were not significantly different from placebo (very low to low quality evidence). However, this should be interpreted cautiously because of the small number of children and the infrequent occurrence of adverse events (Sjoukes et al. 2016). [2018]

Eardrops containing an anaesthetic and an analgesic

  • Eardrops containing an anaesthetic and an analgesic statistically significantly increased the proportion of children with a 50% and a 25% reduction in pain compared with placebo (NNT 5 [range 3 to 16] for 50% pain reduction 10 minutes after having eardrops; low quality evidence). This was based on a systematic review and meta‑analysis of RCTs (Foxlee et al. 2011). These children were aged 3 years and over without eardrum perforation and were also having oral analgesia. [2018]

  • Eardrops containing an anaesthetic and an analgesic statistically significantly reduced antibiotic consumption at day 8 compared with usual care (no or delayed antibiotic prescription; 2.6% compared with 29.0%; moderate quality evidence). There was also a statistically significant reduction in parent‑reported pain scores at day 2 (low quality evidence). This was based on an RCT in children aged 1 to 10 years who did not need immediate antibiotics (Hay et al. 2019). Most children (88%) were also having oral analgesia. [2022]

  • No adverse effects were seen with eardrops containing an anaesthetic and an analgesic, but this was based on very small numbers of children (very low to low quality evidence; Foxlee et al. 2011, Hay et al. 2019). [2018, amended 2022]

Decongestants and antihistamines

  • Overall, decongestants and antihistamines, used alone or in combination, did not improve clinical outcomes in children with acute otitis media who were taking antibiotics (used in 14 out of 15 RCTs; very low to low quality evidence). This was based on a systematic review and meta-analysis of RCTs (Coleman et al. 2008). There was a reduction in the rate of persistent acute otitis media at 2 weeks with a combination of decongestant plus antihistamine compared with placebo (NNT 11 [range 6 to 104]; low quality evidence). However, a subgroup analysis of higher quality studies found no benefit with treatment. [2018]

  • Adverse effects (excluding drowsiness and hyperactivity) were significantly increased with decongestants, but not with antihistamines or a combination of decongestant plus antihistamine, compared with placebo (very low quality evidence). However, there is considerable uncertainty about these results (Coleman et al. 2008). [2018]

Oral corticosteroids

  • Oral prednisolone taken for 5 days did not improve any clinical outcomes in children aged 3 months to 6 years with acute otitis media who were at risk of recurrence (at least 2 previous episodes of acute otitis media), compared with placebo (very low quality evidence). Outcomes included treatment failure during the first 2 weeks, duration of effusion and recurrence. This was based on a small RCT (Chonmaitree et al. 2003). [2018]

  • Adverse effects or discontinuations because of adverse effects did not appear to be significantly different between prednisolone and placebo, although the study was very small and full data were not reported (low quality evidence; Chonmaitree et al. 2003). [2018]

  • Systemic effects (mineralocorticoid and glucocorticoid) may occur with oral corticosteroids, including a range of psychological or behavioural effects (particularly in children; Drug Safety Update on inhaled and intranasal corticosteroids). [2018]

Committee discussion on non-antimicrobial treatments

  • The committee discussed the importance of managing a child's pain and felt that for parents this is the main priority. They agreed that paracetamol or ibuprofen needs to be taken at the right time and at the right dose, with maximum doses being used for severe pain. [2018]

  • Based on evidence and their experience, the committee agreed that paracetamol or ibuprofen should be offered for pain associated with acute otitis media. Parents or carers could be advised to buy paracetamol or ibuprofen over the counter in line with local policies on the prescribing of such medicines. [2018]

  • Based on evidence, the committee agreed that eardrops containing an anaesthetic and an analgesic (in addition to oral analgesics) may reduce antibiotic consumption and relieve pain in children who did not need immediate antibiotics. [2022]

  • They recognised that these eardrops should not be used in children with eardrum perforation or otorrhoea. They discussed the rare adverse effect of methemoglobinemia (associated with topical anaesthetics in very young children). They noted that the age of the children varied in the studies (from 1 year), but that there is no age-based restriction for using the licensed preparation (phenazone with lidocaine [Otigo] eardrops). [2022]

  • They were aware that there is currently only 1 licensed preparation (a prescription only medicine), which was not the preparation used in the studies. However, the committee were content that a class effect would be seen. [2022]

  • There is no direct evidence to support the use of eardrops containing an anaesthetic and an analgesic in children who need immediate antibiotics. The committee agreed that it is uncertain whether these eardrops would provide additional benefit when used with oral analgesia in children having immediate antibiotics. Therefore, they recommended their use only for children who do not need an immediate oral antibiotic. [2022]

  • The committee agreed that evidence does not support using decongestants or antihistamines to help symptoms of acute otitis media. [2018]

  • The committee agreed, based on the evidence, not to make a recommendation on the use of oral corticosteroids to manage acute otitis media in children. [2018]

No antibiotic

  • Acute otitis media is a self-limiting infection of the middle ear. It can be caused by viruses or bacteria, and both are often present at the same time. In most children acute otitis media resolves without treatment. [2018]

  • The most common bacterial causes of acute otitis media are Streptococcus pneumoniae, Haemophilus influenzae, Moraxella catarrhalis and Streptococcus pyogenes (the Canadian Paediatric Society's position statement, 2016). [2018]

  • More common complications of acute otitis media are recurrence of infection, hearing loss (which is usually temporary) and perforated eardrum. However, antibiotics make little difference to the rates of these (see efficacy of antibiotics). [2018]

  • Acute complications of acute otitis media (such as mastoiditis, meningitis, intracranial abscess, sinus thrombosis and facial nerve paralysis) are rare. The incidence of mastoiditis after otitis media is 1.8 per 10,000 episodes after antibiotics compared with 3.8 per 10,000 episodes without antibiotics. This gives a NNT of 4,831 to prevent 1 child from developing mastoiditis (Thompson et al. 2009). [2018]

Efficacy of antibiotics

  • Antibiotics did not significantly reduce pain at 24 hours compared with placebo in children with acute otitis media; around 60% of children in both groups had no pain (high quality evidence). Antibiotics did significantly reduce pain at 2 to 3 days, but the absolute difference was small; 88% of children had no pain in the antibiotic group compared with 84% in the placebo group (NNT 24 [range 15 to 70]; moderate quality evidence). This was based on a systematic review and meta-analysis of RCTs (Venekamp et al. 2015). [2018]

  • Antibiotics significantly reduced the number of children with abnormal tympanometry findings (a surrogate measure for hearing loss) compared with placebo at 2 to 4 weeks, but not at 6 to 8 weeks or 3 months. However, the absolute difference was small; at 2 to 4 weeks, 39% of children had abnormal tympanometry findings with antibiotics compared with 48% with placebo (NNT 12 [range 18 to 21]; low quality evidence; Venekamp et al. 2015). [2018]

  • Antibiotics significantly reduced the number of children with eardrum perforation. However, again the absolute benefits were small with 5% of children having a perforation in the placebo group compared with 2% in the antibiotic group (NNT 33 [range 20 to 100]; moderate quality evidence; Venekamp et al. 2015). [2018]

  • Antibiotics did not reduce the number of children with late recurrence of acute otitis media (which was common in both groups: 18% of children taking antibiotics compared with 20% of children taking placebo, moderate quality evidence; Venekamp et al. 2015). [2018]

  • Antibiotics seem to be more beneficial in 2 pre-defined groups of children, based on subgroup analyses of intervention studies comparing antibiotics with placebo. Firstly, children under 2 years with bilateral acute otitis media, where the NNT was 4 for symptom resolution (low quality evidence). Secondly, children with acute otitis media and otorrhoea (discharge following eardrum perforation), where the NNT was 3 for symptom resolution (moderate quality evidence). This was based on a meta-analysis of individual patient data from RCTs (Rovers et al. 2006). However, the literature search was not designed specifically to identify prognostic evidence. [2018]

  • No systematic reviews or RCTs of topical antibiotics were identified. [2018]

Safety of antibiotics

  • Allergic reactions to penicillins occur in 1 to 10% of people and anaphylactic reactions occur in less than 0.05%. People with a history of atopic allergy (for example, asthma, eczema and hay fever) have a higher risk of anaphylactic reactions to penicillins. People with a history of immediate hypersensitivity to penicillins may also react to cephalosporins and other beta‑lactam antibiotics (BNF information on phenoxymethylpenicillin). See the NICE guideline on drug allergy for more information. [2018]

  • Antibiotic-associated diarrhoea occurs in 2 to 25% of people taking antibiotics, depending on the antibiotic used (NICE Clinical Knowledge Summary on diarrhoea – antibiotic associated). [2018]

  • Adverse events (vomiting, diarrhoea or rash) were significantly increased in children with acute otitis media taking antibiotics compared with those taking placebo (moderate quality evidence). The number needed to harm (NNH) was 13 (range 9 to 25). This was based on a systematic review and meta-analysis of RCTs (Venekamp et al. 2015). [2018]

  • See the summaries of product characteristics for information on contraindications, cautions and adverse effects of individual medicines. [2018]

Committee discussion on no antibiotics
  • Acute otitis media can be caused by viral or bacterial infections, both of which are usually self-limiting and do not routinely need antibiotics. [2018]

  • Based on evidence, the committee agreed that antibiotics make little difference to ear pain or to the rates of more common complications, such as recurrence of infection. The small increased risk of perforation was noted, but 33 children (range 20 to 100) would need to be treated with antibiotics to avoid 1 child experiencing perforation. Antibiotics also made little difference to short-term hearing loss as assessed by the surrogate marker of tympanometry. [2018]

  • More serious complications of acute otitis media, such as mastoiditis, are rare and the NNT with antibiotics to prevent 1 child from developing mastoiditis is approximately 5,000. [2018]

  • The committee acknowledged the recommendation in the NICE guideline on respiratory tract infections (self-limiting): prescribing antibiotics for no antibiotic or a back-up antibiotic prescription for most children with acute otitis media. [2018]

Back-up antibiotics

  • A back-up antibiotic prescription or watchful waiting was as effective as immediate antibiotics in children with acute otitis media for reducing pain at 3 to 7 days (moderate quality evidence). There were also no significant differences between groups for abnormal tympanometry findings (a surrogate measure for hearing loss), eardrum perforation or recurrence of infection (very low to moderate quality evidence). This was based on a systematic review and meta-analysis of RCTs (Venekamp et al. 2015). [2018]

  • A back-up antibiotic prescription was compared with no antibiotics and immediate antibiotics in a systematic review of RCTs (Spurling et al. 2013). In 1 RCT there was no significant difference between back-up antibiotics and no antibiotics for pain or fever on day 3 (very low to low quality evidence). In 1 RCT there was no significant difference between back-up antibiotics and immediate antibiotics for pain on day 3 (moderate quality evidence). [2018]

  • Immediate antibiotics were associated with a significantly increased risk of adverse events (vomiting, diarrhoea or rash) compared with back-up antibiotics or watchful waiting (NNH 8 [range 5 to 19]; moderate quality evidence; Venekamp et al. 2015).

  • The incidence of vomiting or rash was not significantly different with back‑up antibiotics compared with immediate antibiotics (very low quality evidence), but there was significantly less diarrhoea with back‑up antibiotics (NNH 8 [range 5 to 15]; data pooled by NICE; high quality evidence; Spurling et al. 2013). No safety data were available on back‑up antibiotics compared with no antibiotics.

Committee discussion on back-up antibiotics

  • Based on evidence, the committee agreed that no antibiotic prescription or a back-up antibiotic prescription could be considered for most children with acute otitis media.

  • The committee discussed that acute otitis media could have a viral or a bacterial cause, and distinguishing between these is difficult. However, both are usually self-limiting and do not routinely need antibiotics. The committee discussed that a back-up antibiotic prescription may be preferred over no antibiotic in some children, but that prescribers need to weigh up the small clinical benefits from antibiotics against their potential to cause adverse effects.

  • The committee agreed that a back-up antibiotic prescription could be used if symptoms significantly worsen or do not improve within 3 days (by which time most self-limiting infections would be starting to resolve), or if they worsen rapidly or significantly at any time.

  • The committee acknowledged the recommendations in the previous NICE guideline on upper respiratory tract infections that, for acute otitis media, a no antibiotic prescribing strategy or a back-up antibiotic prescribing strategy should be agreed, but that depending on clinical assessment of severity, immediate antibiotics can also be considered for children under 2 years with infection in both ears or children of any age with otorrhoea (discharge following perforation of the eardrum). For these subgroups the committee agreed that an immediate antibiotic prescription could also be considered as an option, because antibiotics may be more likely to be beneficial in these subgroups. The committee discussed that an immediate antibiotic may be preferred over no antibiotic or a back-up antibiotic prescription in some children based on clinical judgement.

  • The committee agreed that immediate antibiotics are important for children who are systemically very unwell, have symptoms or signs of a more serious illness, or are at high risk of serious complications because of pre-existing comorbidity. This includes children with significant heart, lung, renal, liver or neuromuscular disease, immunosuppression, cystic fibrosis, and young children who were born prematurely.

Choice of antibiotic

  • There were no major differences in treatment success between classes of antibiotics, including penicillins, cephalosporins and macrolides for treating uncomplicated acute otitis media in children. There was no difference in treatment success between ampicillin or amoxicillin compared with ceftriaxone; co‑amoxiclav compared with ceftriaxone; co‑amoxiclav compared with azithromycin; or cefaclor compared with azithromycin (low to moderate quality evidence). This was based on a systematic review and meta-analysis of RCTs (Shekelle et al. 2010).

  • Co‑amoxiclav was associated with significantly more adverse events than a cephalosporin (very low to moderate quality evidence) or azithromycin (moderate quality evidence; Shekelle et al. 2010).

  • Shekelle et al. 2010 also considered evidence for treating recurrent or persistent acute otitis media in children. None of the studies found a significant benefit in treatment success for any particular antibiotic (moderate quality evidence). There were 5 individual RCTs that compared different antibiotic treatments: co‑amoxiclav compared with gatifloxacin (2 RCTs), co‑amoxiclav compared with levofloxacin (1 RCT), co‑amoxiclav compared with azithromycin (1 RCT), and cefaclor compared with cefuroxime (1 RCT).

Committee discussion on choice of antibiotic

  • Based on evidence of no major differences in clinical effectiveness between classes of antibiotics, the committee agreed that the choice of antibiotic should largely be driven by minimising the risk of resistance.

  • The committee discussed that, if an antibiotic is needed to treat an infection that is not life-threatening, a narrow-spectrum antibiotic should generally be first choice. Indiscriminate use of broad-spectrum antibiotics creates a selective advantage for bacteria resistant even to these 'last-line' broad-spectrum agents, and also kills normal commensal flora leaving people susceptible to antibiotic-resistant harmful bacteria such as Clostridium difficile. For infections that are not life-threatening, broad-spectrum antibiotics need to be reserved for second-choice treatment when narrow-spectrum antibiotics are ineffective.

  • Based on evidence, their experience and resistance data, the committee agreed to recommend amoxicillin as the first choice because this is current practice for antibiotic treatment in children with acute otitis media, and the risk of resistance is acceptable. The dosage of 125 mg to 500 mg three times a day (based on age) is the usual dose, and was similar to that used in studies in the evidence review. The committee discussed that phenoxymethylpenicillin has a lower risk of resistance than amoxicillin, and microbiologically would be expected to be equivalent. However, medicines adherence is particularly important for children, and acute otitis media most commonly presents in young children. Amoxicillin has a three times a day dosage rather than four times a day for phenoxymethylpenicillin, and the liquid formulation is more palatable.

  • Based on evidence, their experience and resistance data, the committee agreed to recommend clarithromycin as the alternative first-choice antibiotic for use in penicillin allergy or amoxicillin intolerance. In pregnancy, erythromycin was recommended if there is true penicillin allergy. The doses recommended (based on weight and age) are the usual doses for children, and were similar to those used in studies in the evidence review. The committee discussed that there was evidence for another macrolide, azithromycin. However, they agreed not to recommend this because clarithromycin or erythromycin are current practice, and azithromycin should be reserved for more serious infections.

  • The committee discussed the Medicines and Healthcare products Regulatory Agency (MHRA) Public Assessment Report on the safety of macrolide antibiotics in pregnancy. This found that the available evidence is insufficient to confirm with certainty whether there is a small increased risk of birth defects or miscarriage when macrolides are taken in early pregnancy. They agreed with the UK Teratology Information Service monograph on the use of macrolides in pregnancy. They decided that there should be an informed discussion of the potential benefits and harms of treatment. Then, after such a discussion, macrolides can be used if there is a compelling clinical need and there are no suitable alternatives with adequate pregnancy safety data. Erythromycin is the preferred choice if a macrolide is needed during pregnancy, for example, if there is true penicillin allergy and the benefits of antibiotic treatment outweigh the harms. This is because there is more documented experience of its use than for other macrolides. [2022]

  • Based on evidence, their experience and resistance data, the committee agreed to recommend co‑amoxiclav as the second-choice antibiotic for use if symptoms worsen on a first-choice antibiotic taken for at least 2 to 3 days. This broad-spectrum treatment combines a penicillin (amoxicillin) with a beta-lactamase inhibitor, making it active against beta-lactamase-producing bacteria that are resistant to amoxicillin alone. People who do not respond to amoxicillin may be more likely to have an infection that is resistant to it. The dosage of 0.25 ml/kg of 125/31 suspension to 250/125 mg or 500/125 mg three times a day (based on weight and age) is the usual dose for children, and was similar to that used in studies in the evidence review. [2018]

Antibiotic course length

  • A short course of antibiotics (more than 48 hours but less than 7 days) was associated with significantly higher treatment failure at 8 to 19 days, or 1 month or less, compared with a long course (7 days or longer). Treatment failure (defined as a lack of clinical resolution, relapse or recurrence of acute otitis media within 1 month of starting treatment) occurred in 18.0% of the short-course group compared with 14.4% of the long-course group at 8 to 19 days (NNT 28 [range 17 to 77]; very low quality evidence), and in 20.5% of the short-course group compared with 17.5% of the long-course group at 1 month or less (NNT 34 [range 20 to 124]; low quality evidence). However, there was no difference in treatment failure between short and long courses at other time points. This was based on a systematic review and meta‑analysis of RCTs (Kozyrskyj et al. 2010). [2018]

  • There were significantly fewer gastrointestinal adverse events with a short course of antibiotics (more than 48 hours but less than 7 days) compared with a long course (7 days or longer; very low quality evidence). However, this result was based on the reported odds ratio and was not statistically significant when the relative risk was calculated (Kozyrskyj et al. 2010). [2018]

Committee discussions on antibiotic course length

  • The committee agreed that, when an antibiotic is appropriate, the shortest course that is likely to be effective should be prescribed to minimise the risk of antimicrobial resistance. [2018]

  • Based on evidence, their experience and resistance data, the committee agreed that a 5‑ to 7‑day course of all the recommended antibiotics was sufficient to treat acute otitis media in children. This takes into account both the evidence for clinical effectiveness and the evidence for safety and tolerability of antibiotics, and minimises the risk of resistance. Studies on the use of specific antibiotics to treat acute otitis media sometimes had longer course lengths than 7 days. [2018]

  • The committee noted that no studies were identified that directly compared a 5‑day course of antibiotics with a 7‑day course. [2018]

  • Based on evidence, the committee recognised that more children may have treatment failure with an antibiotic course of less than 7 days compared with a course of 7 days or more. However, the absolute difference is small. At 8 to 19 days, 82% of children taking antibiotics for less than 7 days were better, compared with 86% of those taking antibiotics for 7 days or more. They agreed that, if a decision to prescribe an antibiotic is made, a 5‑day course may be sufficient for many children, reserving 7‑day courses for those with a clinical assessment of more severe or recurrent infection. [2018]

Antibiotic dose frequency

  • Once or twice daily dosing of amoxicillin or co‑amoxiclav was as effective as three times a day dosing for clinical cure rates at the end of antibiotic treatment (high quality evidence). The duration of treatment was 10 days in most studies, and the dose of amoxicillin or co‑amoxiclav varied. There were no significant differences in the rates of recurrence (very low to low quality evidence), adverse effects (very low to low quality evidence) and adherence (high quality evidence). This was based on a systematic review and meta‑analysis of RCTs (Thanaviratananich et al. 2013). [2018]

Committee discussions on antibiotic dose frequency

  • The committee discussed the evidence for once or twice daily dosing of amoxicillin and co‑amoxiclav, but it is unknown if this would have a detrimental effect on the risk of resistance to these antibiotics. The evidence supporting once or twice daily dosing is for different doses and longer treatment durations. This goes against the general principle of antimicrobial stewardship to prescribe the shortest course that is effective. [2018]

  • The committee agreed that, when prescribing amoxicillin or co‑amoxiclav, a dosing frequency of three times a day should be prescribed, as is current practice. [2018]

See the full evidence review for more information.

  • National Institute for Health and Care Excellence (NICE)