OBJECTIVE To assess the ability of ED tympanometry to identify those children with acute otitis media who are at greater risk for treatment failure, recurrent infection, or persistent middle ear effusion. METHODS A prospective observational study was performed in the EDs of a county hospital and a university medical center. One hundred seventeen children, aged 5 months to 5 years, with acute otitis media and abnormal tympanometric patterns were categorized into three groups--low compliance, high pressure, and low pressure--based on the initial patterns. All the children received oral amoxicillin for ten days. They had repeat tympanometry and otoscopy at 14- and 60-day follow-up visits, and were followed to determine clinical outcomes. RESULTS Initial treatment success was equivalent for all groups: low compliance 93% (95% confidence interval 81%-97%), high pressure 91% (95% CI 71%-98%), low pressure 89% (95% CI 59%-99%). Rates of relapse, reinfection, 90-day non-recurrence, and persistent otoscopic and tympanometric middle ear effusion did not differ significantly between groups. CONCLUSIONS ED tympanometry of children with acute otitis media is unable to provide prognostic information regarding treatment failure, recurrent infection, or persistent middle ear effusion.