Structured clinical question

Do children with OME [population] have an increased risk of AR [outcome] than children without OME [comparison]?

Search strategy and outcome

Our search strategy (extended to 2 August 2003) was:

• Cochrane Database of Systematic Reviews using: “otitis media AND allergy”; 13 references (none relevant).

• Medline, via Pubmed: “otitis media with effusion AND allergic rhinitis”; 62 references (four relevant).

See table 1.

Commentary

The full text of two relevant studies^3,4 was not accessible to us (one was published in Japanese, the other in Turkish); however, the abstract furnished sufficient details for a summary evaluation of their validity and utility for our question.

The studies that we examined in full text^1,2 showed marked difference in the prevalence of AR in children with OME: 16.3% versus 89%. Trying to explain this discrepancy, it can be noted that the study of Alles and colleagues² is affected by some methodological imperfections that seriously compromise its validity. It lacks a well defined control group and the study definitions of AR and OME are weak. For AR, neither the appearance of the symptoms after exposure to an allergen nor the demonstration, necessarily, of sensitisation to an allergen through measurement of the specific IgE is required. Even the definition of OME was not strong: an unconfirmed history of OME was sufficient for enrolment. The prevalence of the AR in children with OME in the study of Caffarelli and colleagues¹ gives the more reliable estimates; because of the fact that their study is prospective, and the authors have adopted rigorous diagnostic criteria for both the illness studied (AR and OME), have included an adequate control population, and have enrolled a sufficient number of children.

The prospective design of the study reduces the recall bias and the possibility of differences in the management of the patients. A rigorous definition of allergic rhinitis allows avoiding its overdiagnosis and the inclusion of patients with non-allergic rhinitis. OME was defined prospectively with tympanometry performed on all the patients (to define cases and controls). The presence of a control group allows us to quantify the parameter we consider the most interesting—that is, the difference (absolute risk increase) in the prevalence of the allergic rhinitis between the children with OME and the ones without. And finally, the presence of a large sample makes the estimate of the absolute risk increase more accurate, tightening its confidence interval.

The duration of OME is an important variable in management decisions: studies of children in day care note that many will have brief periods of time (one or several days) with OME that spontaneously clears; in contrast other children will have OME for months. Only for the latter children would any intervention have the potential to be useful.

However, the studies on the efficacy of antiallergic therapy in the treatment of children with OME are few, methodologically weak, and inconclusive. “Irrespective of the theoretical mechanism, the relation between allergy and otitis media with effusion will remain controversial until well controlled clinical studies are conducted documenting that in select populations antiallergy therapy is efficacious in preventing or limiting the duration of otitis media with effusion”.⁵ Today, the situation is unchanged.