Is secretory otitis media a single disease entity?

The outcome of 809 children who had middle ear effusion (MEE) was correlated with their age and medical history and the bacteriologic and cytologic findings of the MEE. Three groups emerged. Group A (n = 384) had a medical history of relatively recent acute otitis media (AOM) with a peak prevalence of 2.6 years of age, and its MEE was hypercellular (mostly polymorphonuclear leukocytes) with a 24% rate of positive bacteriologic culture. In contrast, the peak prevalence of group B (n = 280) was approximately 5 years of age. Its MEE followed no otologic medical history and was bacteriologically sterile with relatively few cells (mostly lymphocytes typical of viral infections). Group AB (n = 145) resembled group B except that they had a history of AOM some years before hearing loss onset. Altogether, the clinical features of group A are statistically distinguishable from those of groups B and AB in most respects. These findings explain the bimodal peak prevalence distribution that was found in many available epidemiological studies of secretory otitis media (SOM) and that was also seen in our data. It appears that SOM should neither be termed nor treated as an otitis media, but as a sequela of either bacterial AOM (group A) or of insidious, asymptomatic, probably viral otitis media (groups B and AB). Our data do not support antibiotic treatment for SOM (otitis media with effusion)--especially not in chronic cases.