Level 1: Likely reliable evidence

Various guidelines suggest analgesics either as an alternative to antibiotics for the treatment of uncomplicated acute otitis media (AOM) in children or as the preferred option. Although antibiotics have been shown to reduce symptoms, many studies and meta-analyses have suggested that the benefit of such treatment may be small, since symptoms often resolve on their own. One criticism of earlier trials is that the criteria for AOM may have been loosely defined, including children with only mild illness. In two new randomized trials, amoxicillin-clavulanate was compared vs. placebo in populations with strict criteria for AOM diagnosis.

In the first trial, 322 children aged 6-35 months with AOM were randomized to amoxicillin 40 mg/kg/day plus clavulanate 5.7 mg/kg/day orally vs. placebo for seven days (N Engl J Med. 2011;364:116-26). For AOM diagnosis, children had to have symptoms, acute inflammatory signs in the tympanic membrane, and otoscopic indications. Treatment failure was defined as a composite of several factors: no symptom improvement after three days or no otoscopic improvement after eight days, worsening condition or tympanic membrane perforation at any time, infection requiring additional antimicrobial treatment or discontinuation of the study drug.

Failure occurred in 18.6% of the amoxicillin-clavulanate group vs. 44.9% of the controls (P <0.001, NNT 4). Rescue medication was needed in 6.8% vs. 33.5% (P <0.001, NNT 4). There were no significant differences in the use of analgesics or antipyretics. The amoxicillin-clavulanate group had significantly higher rates of diarrhea and eczema.

In a second trial with similar diagnostic criteria, 291 children aged 6-23 months were randomized to amoxicillin 90 mg/kg plus clavulanate 6.4 mg/kg orally daily vs. placebo for 10 days (N Engl J Med. 2011;364:105-115; available at www.nejm.org/doi/full/10.1056/NEJMoa0912254, accessed February 15, 2011). Amoxicillin-clavulanate was associated with greater rates of symptom resolution at two, four, and seven days (P=0.04 overall, at seven days 67% vs. 53%, NNT 8) and with lower rates of treatment failure at four to five days (4% vs. 23%, P <0.001, NNT 6) and at 10 to 12 days (16% vs. 51%, P <0.001, NNT 3). Rates of diarrhea and diaper-area dermatitis were significantly higher in the amoxicillin group.