Do children's ear infections need antibiotics, and what helps at home?

An ear infection (acute otitis media) is infected fluid behind the eardrum, usually after a cold. In high-income countries most cases clear up without complications².

What antibiotics do. A Cochrane review of 13 placebo-controlled trials (3,401 children) found antibiotics did not reduce pain at 24 hours. They cut the number still in pain at 2 to 3 days by almost a third (about 1 child in 20 benefits), lowered eardrum perforations and infection in the other ear, and caused vomiting, diarrhea or rash in 1 in 14². Mastoiditis, the feared complication, is rare: about 4,800 episodes would need antibiotics to prevent one case⁸.

Why young children differ. In two trials of children under 3 with strictly diagnosed infections, antibiotics cut treatment failure from 51% to 16%⁵ and from 45% to 19%⁶. The AAP therefore advises antibiotics for severe illness (fever of 39 °C or higher, moderate or severe pain, or pain for 48 hours or more) and for infections in both ears at 6 to 23 months. For milder cases it advises a shared choice between antibiotics and observation¹.

Waiting is workable. In an emergency-department trial, 62% of parents given a wait-and-see prescription never filled it, with no more fever or ear pain afterwards⁹.

At home. Pain relief is the mainstay, with or without antibiotics. The AAP found no controlled studies of home remedies and says heat or cold may have limited effect¹. An herbal drop (garlic, mullein, calendula and St John's wort in olive oil) eased pain as well as anesthetic drops in 103 children aged 6 to 18, but there was no placebo group⁷. See mullein for the herb's wider evidence.

Prevention. Exclusive breastfeeding for 6 months went with about 43% fewer infections³. Living with a smoker raised the odds by 37% to 62%⁴. Pacifier use was tied to recurrent infections¹⁰. Pneumococcal vaccine mainly cuts pneumococcal ear infections¹¹.

Recurrent infections. The AAP advises against preventive antibiotics and says tubes may be offered after 3 infections in 6 months or 4 in a year¹. Trials show about one fewer infection over six months with tubes, with low-quality evidence¹².

Antibiotics versus placebo or waiting

The 2023 Cochrane update pooled 13 trials from high-income countries (3,401 children). Antibiotics did not reduce pain at 24 hours (risk ratio 0.89, 95% CI 0.78 to 1.01; high-certainty evidence). At 2 to 3 days almost a third fewer children had pain (RR 0.71, 95% CI 0.58 to 0.88; number needed to treat 20), and at 10 to 12 days fewer still (RR 0.33, one trial). Adverse events rose (RR 1.38, number needed to harm 14). Antibiotics slightly reduced eardrum perforation (RR 0.43), halved infections in the other ear (RR 0.49), and did not prevent later recurrences. Serious complications were rare and similar in both arms. Comparing immediate antibiotics with expectant observation across 6 trials (1,556 children), adverse events were again higher (RR 1.87, number needed to harm 10). The authors concluded observation seems justified for most children with mild disease, with emphasis on analgesia².

Young children: the counterweight

Hoberman randomized 291 children aged 6 to 23 months, diagnosed by stringent criteria, to 10 days of amoxicillin-clavulanate or placebo. Sustained symptom resolution by day 7 was 67% versus 53%, and clinical failure at day 10 to 12 was 16% versus 51% (P < 0.001); one placebo child developed mastoiditis, and diarrhea and diaper dermatitis were more common with antibiotics⁵. Tähtinen randomized 319 children aged 6 to 35 months to 7 days of the same drug or placebo. Treatment failure was 18.6% versus 44.9% (hazard ratio 0.38, 95% CI 0.25 to 0.59), rescue treatment 6.8% versus 33.5%, and diarrhea 47.8% versus 26.6%⁶. These trials explain why the pooled Cochrane benefit looks modest while guidelines favor antibiotics in the youngest children and severe disease.

Wait-and-see prescribing and mastoiditis

Spiro randomized 283 children aged 6 months to 12 years in an emergency department to a wait-and-see or standard prescription; all got ibuprofen and analgesic ear drops. Parents in the wait-and-see group left the prescription unfilled 62% of the time versus 13%, with no significant difference in later fever, ear pain or unscheduled visits; fever (RR 2.95) and ear pain (RR 1.62) predicted who filled it⁹. In a UK database of 2.6 million children, mastoiditis stayed at about 1.2 per 10,000 child-years from 1990 to 2006. After an ear infection the risk was 1.8 per 10,000 episodes with antibiotics versus 3.8 without, so 4,831 episodes would need treating to prevent one case, and only about a third of children with mastoiditis had seen a doctor for otitis media beforehand⁸.

Home remedies and herbal drops

Sarrell's trial of 103 children aged 6 to 18 compared an herbal extract (Allium sativum, Verbascum thapsus, calendula and Hypericum perforatum in olive oil) with anesthetic drops; pain scores improved significantly over 3 days and the groups were comparable⁷. Limits: no placebo or no-drop arm, a commercial product, and no children under 6. The AAP adds that a Cochrane review of topical analgesia found limited evidence of benefit at 30 minutes and could not separate natural recovery, placebo effect, the soothing effect of any liquid and the drops themselves¹.

Risk factors

Bowatte's meta-analysis of 24 studies found exclusive breastfeeding for 6 months cut the odds of an ear infection in the first 2 years (OR 0.57, 95% CI 0.44 to 0.75), and any breastfeeding helped less (OR 0.67)³. Jones's review of 61 studies found maternal postnatal smoking raised the odds of middle ear disease (OR 1.62, 95% CI 1.33 to 1.97), any household smoker 1.37, and surgery for middle ear disease even more⁴. Rovers followed 495 children aged 0 to 4: pacifier use was not clearly linked to a first infection (OR 1.3, 95% CI 0.9 to 1.9) but was to recurrent infections (OR 1.9, 95% CI 1.1 to 3.2)¹⁰. A Cochrane review of 11 trials (60,733 children) found early-infancy pneumococcal vaccines cut pneumococcal ear infections by 20% to 25% (PCV7) and about 50% (PHiD-CV10 and CV11), but all-cause infections by only about 6% in the CRM197-PCV7 trials in low-risk infants¹¹.

Ear tubes

Five small trials (805 children, all run before pneumococcal vaccination) compared tubes with monitoring or medication. Tubes cut recurrences, for example no recurrence at 6 months in 46% versus 5% in one 95-child trial, but the benefit faded by 12 months (48% versus 34%) and the evidence was low quality; persistent perforation was uncommon¹².

Bottom line (confidence 4 of 5)

Pain control and, for mild cases in older children, monitored waiting are well supported; young children and severe cases do benefit from antibiotics. Herbal drops and heat remain unproven comfort measures. Follow your pediatrician's advice, especially for babies under 6 months.