Acute Otitis Media Treatment in Young Children
Weighing immediate antibiotics against watchful waiting — and deciding how long to treat
One-Minute Takeaway
- For most children with acute otitis media, high-certainty evidence shows antibiotics do not relieve pain within 24 hours (RR 0.89, 95% CI 0.78–1.01) and cut early pain only modestly (NNTB 20 at 2–3 days).
- Benefit concentrates in defined groups: the number needed to treat falls to 3 with ear discharge and 4 in under-2s with both ears affected.
- Two rigorous placebo-controlled trials in young children roughly halved clinical failure with antibiotics (Tähtinen: hazard ratio 0.38, 95% CI 0.25–0.59).
- In children under 2, a 5-day course was inferior to 10 days (failure 34% vs 16%); shorter courses are safer to weigh in older children.
- Guidelines endorse watchful waiting with a safety-net plan for older children with mild, one-sided disease.
Why Acute Otitis Media Matters
Acute otitis media is among the most common infections of early childhood and the leading reason children receive antibiotics. Most children have at least one episode before school age.
Prescribing habits diverge sharply. Antibiotic use for the condition ranges from about 56% of cases in the Netherlands to roughly 95% in the United States, Canada, and Australia.
That variation matters because most episodes in high-income settings settle on their own. Pooled trial data indicate spontaneous resolution in the large majority of children, without antibiotics and without complications.
The clinical task, then, is not simply whether antibiotics work, but for which child, and for how long. Getting that balance right protects the children who genuinely benefit while limiting adverse effects and antimicrobial resistance in those who do not.
The Clinical Problem in Acute Otitis Media
The core difficulty in acute otitis media is separating the child who needs an antibiotic from the many who will recover without one. Overtreatment carries real costs, and undertreatment risks a protracted course in vulnerable infants.
Antibiotics are not harmless. In the most recent Cochrane synthesis, they increased vomiting, diarrhoea, or rash compared with placebo (RR 1.38, 95% CI 1.16–1.63; high-certainty evidence), corresponding to one additional child harmed for every 14 treated.
Diagnosis adds further uncertainty. A bulging tympanic membrane is required to distinguish true acute otitis media from otitis media with effusion, a common viral-associated state that does not respond to antibiotics.
Two decisions follow from a confident diagnosis. First, treat immediately or observe with a safety net. Second, if treating, choose a standard or a shortened course. The sections below compare the evidence behind each option.
Available Approaches
Three management approaches dominate current practice: immediate antibiotic therapy, watchful waiting with a delayed prescription, and shortened-course antibiotics for children who are treated. The comparative evidence base is summarised below, then examined approach by approach.
| Study / Source | Design | N | Population | Comparison | Key result |
|---|---|---|---|---|---|
| Venekamp 2023 (Cochrane) V | Systematic review / meta-analysis of RCTs | 2,320 (7 RCTs, pain outcome) | Children with AOM, high-income settings | Antibiotics vs placebo | No pain relief at 24 h (RR 0.89, 95% CI 0.78–1.01); pain at 2–3 d RR 0.71 (0.58–0.88), NNTB 20; harms RR 1.38 (1.16–1.63), NNTH 14 |
| Rovers 2006 (Lancet) V | Individual patient-data meta-analysis (6 RCTs) | 1,643 | Children 6 mo–12 y with AOM | Antibiotics vs control, by subgroup | Otorrhoea: NNT 3 (rate difference −36%, −53 to −19); under-2s with bilateral AOM: NNT 4 (−25%, −36 to −14); no otorrhoea: NNT 8 |
| Tähtinen 2011 (NEJM) V | Double-blind RCT | 319 | 6–35 mo, strict criteria | Amoxicillin-clavulanate vs placebo × 7 d | Treatment failure 18.6% vs 44.9%; progression to failure HR 0.38 (0.25–0.59); rescue treatment 6.8% vs 33.5% |
| Hoberman 2011 (NEJM) V | Double-blind RCT | 291 | 6–23 mo, stringent criteria | Amoxicillin-clavulanate vs placebo × 10 d | Clinical failure by day 10–12: 16% vs 51%; absolute reduction 35% (95% CI 25–45); NNT 3 |
| Hoberman 2016 (NEJM) V | Double-blind RCT (noninferiority) | 520 | 6–23 mo | 5-day vs 10-day amoxicillin-clavulanate | Clinical failure 34% vs 16%; difference 17 percentage points (95% CI 9–25); no reduction in adverse events or resistance with the shorter course |
| Kim 2024 (World J Pediatr) V | Systematic review / network meta-analysis | 89 RCTs | Children and adolescents with AOM | 7-day vs 10-day first-line regimens | 7 days noninferior to 10 days for amoxicillin (RR 0.92, 0.82–1.03) and amoxicillin-clavulanate (RR 1.11, 0.96–1.28) — except in children ≤2 years |
All figures verified against the primary source or a systematic review retrieved during preparation. [V] = verified. Effect sizes are shown with 95% confidence intervals.
Immediate Antibiotic Therapy
Two placebo-controlled trials in young children, both using strict otoscopic criteria, provide the strongest case for immediate treatment. In the Hoberman 2011 trial, amoxicillin-clavulanate reduced clinical failure by day 10–12 from 51% to 16% (absolute reduction 35%, 95% CI 25–45; NNT 3).
The Tähtinen 2011 trial found a similar pattern. Treatment failure fell from 44.9% to 18.6%, and the rate of progression to failure was reduced by 62% (HR 0.38, 95% CI 0.25–0.59). The need for rescue treatment dropped from 33.5% to 6.8%.
These benefits are real but come with a caveat. Both trials also recorded more diarrhoea and diaper dermatitis with active treatment, and one child in the Hoberman placebo group developed mastoiditis. Amoxicillin remains the first-line agent unless the child took it within 30 days, has purulent conjunctivitis, or is penicillin-allergic.
Clinical Pearl
The striking difference in these two trials was driven partly by high failure rates in the placebo groups, reflecting stringent diagnostic entry criteria. Where the diagnosis is certain and the child is young, immediate treatment has the clearest support.
Watchful Waiting and Delayed Prescribing
Watchful waiting means active monitoring with symptom relief, holding antibiotics unless the child worsens or fails to improve within 48–72 hours. A safety-net (delayed) prescription lets a caregiver start treatment if that happens, without a return visit.
The strategy rests on how modest average benefit is. High-certainty evidence shows antibiotics do not reduce pain at 24 hours (RR 0.89, 95% CI 0.78–1.01) and reduce it only slightly at 2–3 days (RR 0.71, 0.58–0.88; NNTB 20). Most children recover regardless.
Individual patient-data pooling clarifies who benefits enough to skip observation. Antibiotics help most in children with ear discharge (NNT 3) and in under-2s with bilateral disease (NNT 4), and least in children without discharge (NNT 8). Age alone was not a significant modifier.
The 2013 American Academy of Pediatrics guideline builds on this. It offers observation, through joint decision-making with caregivers, for unilateral non-severe acute otitis media in children 6–23 months and for non-severe disease in children 2 years and older, provided follow-up is assured.
Shortened-Course Treatment
Duration is the second decision, and here age is decisive. In the Hoberman 2016 noninferiority trial, children aged 6–23 months treated for 5 days had more clinical failure than those treated for 10 days (34% vs 16%; difference 17 percentage points, 95% CI 9–25).
Shortening the course in these young children did not deliver the hoped-for trade-off. Rates of adverse events, recurrence, and resistant nasopharyngeal colonisation were not lower with 5 days than with 10.
Older children appear different. A network meta-analysis of 89 trials found a 7-day course noninferior to 10 days for amoxicillin (RR 0.92, 95% CI 0.82–1.03) and amoxicillin-clavulanate (RR 1.11, 0.96–1.28), with the exception of children under 2. An earlier Cochrane review of short versus long courses found a small increase in one-month treatment failure with shorter regimens (OR 1.34, 95% CI 1.15–1.55; absolute difference about 3%, NNT 33), alongside fewer gastrointestinal effects.
Comparative Analysis
Placed side by side, the three approaches map onto the child in front of the clinician rather than onto a single universal rule. The evidence favours matching intensity of treatment to risk.
For the youngest and most clearly affected — infants under 2 with bilateral disease, or any child with otorrhoea — immediate antibiotics carry the strongest support, with numbers needed to treat of 3 to 4. For these children, a full 10-day course is also better supported than a shortened one.
For older children with mild, unilateral disease, the calculus shifts. Average benefit is small, spontaneous resolution is the norm, and watchful waiting with a safety-net prescription is a reasonable, guideline-endorsed default. When these children are treated, a 7-day course appears to give up little.
The consistent thread across trials and reviews is that harms are stable while benefits vary. Adverse events run at roughly one additional affected child per 14 treated regardless of who is treated, which is why targeting matters.
Quality and Consistency of the Evidence
The evidence base is unusually strong for a common paediatric condition, anchored by multiple double-blind randomized trials and Cochrane reviews. Certainty nonetheless varies by outcome.
- Antibiotics versus placebo for early pain — High certainty
Multiple RCTs and a current Cochrane review agree: no benefit at 24 hours, modest benefit at 2–3 days. Findings are consistent across settings. - Reduction in clinical failure in young children — High certainty
Two independent placebo-controlled trials using strict criteria show concordant, sizeable reductions in treatment failure. - Adverse events of antibiotics — High certainty
Increased vomiting, diarrhoea, and rash are precisely estimated and reproducible (RR 1.38, 95% CI 1.16–1.63). - Which subgroups benefit most — Moderate certainty
Individual patient-data pooling is robust for otorrhoea and bilateral under-2s, but subgroups were limited and defined post hoc. - Shorter courses in older children — Moderate certainty
Network and pairwise meta-analyses suggest 7 days suffices beyond infancy, but head-to-head 5-day data outside young children are sparse.
What the Evidence Does Not Show
Specific gaps, not a call for generic further study
The trials measured short-term clinical and symptomatic outcomes. They do not establish whether treatment choice changes long-term hearing, speech, or language development.
Several concrete questions remain open. The 5-day data that show inferiority apply specifically to children under 2; whether 5 days is adequate in older children has not been tested with the same rigour.
Head-to-head comparison of a delayed prescription against genuine no-antibiotic observation is limited, so the incremental value of the safety-net prescription itself is uncertain. Diagnostic accuracy also varies between clinicians, and much of the trial evidence predates widespread pneumococcal conjugate vaccination, which has shifted the organisms and rates involved.
Finally, the evidence does not identify a single threshold that reliably separates children who need antibiotics from those who do not, beyond the broad markers of age, laterality, discharge, and severity.
Practical Implications
The evidence supports pain assessment and analgesia for every child with acute otitis media, whether or not antibiotics are prescribed. Symptom relief is the one intervention that helps essentially all children.
It supports immediate antibiotics where benefit is concentrated: children with otorrhoea, under-2s with bilateral disease, and any child with severe signs such as moderate-to-severe otalgia, otalgia beyond 48 hours, or a temperature of 39°C or higher.
It supports offering watchful waiting, with a mechanism for follow-up or a safety-net prescription, to older children and to those with mild, one-sided disease. The plan should trigger antibiotics if the child worsens or fails to improve within 48–72 hours.
On duration, the evidence supports a full 10-day course for children under 2 and makes a 7-day course a reasonable option in older children. Amoxicillin remains first-line for most, reserved to a beta-lactamase–stable alternative after recent use, purulent conjunctivitis, or allergy.
Evidence Grade and Bottom Line
Strong evidence supports a targeted approach to acute otitis media rather than a uniform one. Antibiotics offer little to the average child in the first day and a modest early benefit thereafter, while reliably adding adverse effects.
The children who benefit enough to warrant immediate treatment are identifiable: the very young with bilateral disease, those with ear discharge, and those with severe symptoms. For them, a standard 10-day course is best supported.
For everyone else, watchful waiting with assured follow-up is a defensible default, and a 7-day course is reasonable when antibiotics are used beyond infancy. Matching treatment to the child, and pairing every decision with good pain control, is what the current evidence most clearly endorses.
Article Information and References
- Venekamp RP, Sanders SL, Glasziou PP, Rovers MM. Antibiotics for acute otitis media in children. Cochrane Database Syst Rev. 2023;11(11):CD000219. DOI: https://doi.org/10.1002/14651858.CD000219.pub5
- Rovers MM, Glasziou P, Appelman CL, et al. Antibiotics for acute otitis media: a meta-analysis with individual patient data. Lancet. 2006;368(9545):1429–1435. DOI: https://doi.org/10.1016/S0140-6736(06)69606-2
- Kim MS, Kim JH, Ryu S, et al. Comparative efficacy and optimal duration of first-line antibiotic regimens for acute otitis media in children and adolescents: a systematic review and network meta-analysis of 89 randomized clinical trials. World J Pediatr. 2024;20(3):219–229. DOI: https://doi.org/10.1007/s12519-023-00716-8
- Kozyrskyj A, Klassen TP, Moffatt M, Harvey K. Short-course antibiotics for acute otitis media. Cochrane Database Syst Rev. 2010;(9):CD001095. DOI: https://doi.org/10.1002/14651858.CD001095.pub2
- Tähtinen PA, Laine MK, Huovinen P, Jalava J, Ruuskanen O, Ruohola A. A placebo-controlled trial of antimicrobial treatment for acute otitis media. N Engl J Med. 2011;364(2):116–126. DOI: https://doi.org/10.1056/NEJMoa1007174
- Hoberman A, Paradise JL, Rockette HE, et al. Treatment of acute otitis media in children under 2 years of age. N Engl J Med. 2011;364(2):105–115. DOI: https://doi.org/10.1056/NEJMoa0912254
- Hoberman A, Paradise JL, Rockette HE, et al. Shortened antimicrobial treatment for acute otitis media in young children. N Engl J Med. 2016;375(25):2446–2456. DOI: https://doi.org/10.1056/NEJMoa1606043
- Lieberthal AS, Carroll AE, Chonmaitree T, et al. The diagnosis and management of acute otitis media. Pediatrics. 2013;131(3):e964–e999. DOI: https://doi.org/10.1542/peds.2012-3488