Diagnostic inaccuracy and subject exclusions render placebo and observational studies of acute otitis media inconclusive.

BACKGROUND

Diagnostic accuracy and appropriate inclusion/exclusion criteria representative of children at greatest risk is of paramount importance in trials to evaluate placebo or observation as an option for acute otitis media (AOM) management.

METHODS

Twelve observational studies spanning the time frame 1958-2005 and 13 natural history studies spanning the time frame 1968-2006 were evaluated for the diagnostic criteria, inclusion criteria, and exclusion criteria applied within the study design.

RESULTS

Although a bulging or full tympanic membrane (TM) with effusion is the best indication of a diagnosis of bacterial AOM based on tympanocentesis findings, few observational and natural history studies required a bulging TM. Examination of subject inclusion criteria showed that many subjects did not have AOM but rather had no middle ear disease at all or they had otitis media with effusion. Exclusion criteria of subjects were also remarkable. Frequently children <2 years old were excluded; mean age among the studies reflected an older age group, unlike the true epidemiology of AOM. Otitis prone children, those with severe disease, with a bulging TM, with fever, with a definite need for antibiotics, with recent antibiotic treatment, with recent AOM, or with perforation of the TM were often excluded.

CONCLUSIONS

Guidelines and some authorities have overlooked or discounted the importance of the issues of inaccurate diagnosis on study entry, broad inclusion criteria, and the creation of bias in exclusion criteria among placebo/natural history trials in AOM. The current data favoring observation of children with AOM should be reconsidered until better studies are conducted.