Statistics from Altmetric.com
Q Are antibiotics effective in children with acute otitis media?
Medline, EMBASE/Excerpta Medica, Current Contents, and the Cochrane Controlled Trials Register (to March 2003).
Study selection and assessment:
randomised controlled trials (RCTs) that compared antimicrobial drugs with placebo in children with otitis media (without tympanostomy tubes). Methodological quality was assessed for randomisation method, blinding, control of selection bias after treatment assignment, and outcome assessment.
pain (severity and duration), hearing (tympanometry or audiometry test results), progression of symptoms (eg, vomiting, diarrhoea, or rash; and recurrence), and complications.
10 RCTs met the selection criteria, of which 8 included patient relevant outcomes. Study quality was generally high. Groups did not differ for resolution of pain at 24 hours, but fewer patients who received antibiotics still had pain at 2–7 days (table). Of the trials that assessed hearing, no difference between groups was seen at 1 month (3 RCTs) or 3 months (2 RCTs). Vomiting, diarrhoea, or rash was increased in the antibiotics group (table). Groups did not differ for recurrence (table). Complications were rare in both groups (1 patient with mastoiditis).
In children with acute otitis media, antibiotics are effective for resolution of pain at 2–7 days, but the number needed to treat is 15 children in order to reduce pain in 1 child.
The excellent review by Glasziou et al updates a previous review on this subject. The authors acknowledge 2 issues that immediately come to mind for the practitioner.
Firstly, the inclusion criteria for the studies varied, and this could have had a substantial effect on the meta-analysis. Acute otitis media has often been described as being overdiagnosed by practitioners, and the criteria for diagnosis are still debated. Appelman et al found that signs of tympanic inflammation did not predict the course of acute otitis media.1 Pneumatic otoscopy was proposed as an alternative criteria, but a review by Preston showed that pneumatic otoscopy predicted only serous otitis and not acute otitis media.2 Therefore, accurate diagnosis is still an issue and confounds conclusions about outcomes. Including children with minimal symptoms in trials might blunt the effect seen when treating those with genuine acute otitis media.
Secondly, when children present with red ear and pain, clinicians are presented with the dilemma of who to treat to reduce pain, while trying to avoid the occasional serious complications of hearing loss and mastoiditis, as well as overprescription of antibiotics. Little et al found that children with fever and vomiting were more likely to have poor outcomes and benefited from antibiotic treatment.3 Another study by Little et al proposed waiting 72 hours to treat.4 Either approach seems like a sensible solution based on the quality of current evidence.
For correspondence: Professor P P Glasziou, University of Oxford, Oxford, UK.
Source of funding: not stated.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.