Treatment for otitis media
==========================

* Nicole Le Saux
* Isabelle Gaboury
* David Moher

Barry Pless asks why a noninferiority trial was chosen over an equivalence trial for our study of the treatment of otitis media.1 There are subtle but important differences between these types of trials in terms of both sample size requirements and interpretation. We were not aiming to demonstrate statistically significant superiority of amoxicillin over placebo, as this has already been well established. Our objective was to show that waiting to treat (which we simulated by using placebo) would not be substantially worse (within a 10% clinically acceptable difference in failure rates compared with amoxicillin) than giving amoxicillin to children presenting with acute otitis media. In recent terminology, this is known as a noninferiority trial. By contrast, an equivalence trial seeks to demonstrate that the outcomes of 2 interventions are not substantially different, without prespecifying which intervention will lead to better results. Hence, for our question, a noninferiority trial was appropriate.

In response to Mathieu Lemaire, we found the article by Gomberg-Maitland and associates2 off the mark. The context of the Lemaire quotation was a 3-arm study (placebo, active control and experimental treatment), which was not the design of our study. Similarly, Lemaire's last remark (following the citation from Chen and colleagues3) is in reference to noninferiority trials with active control, which was not part of the design of our trial.

Brian Blakley brings up an important clinical point. Within the clinical definition for acute otitis media one has to meet the “time” criterion of abrupt onset of signs and symptoms of middle ear inflammation and middle ear effusion. This presentation is distinct from otitis media with effusion, which is a subacute or chronic problem, does not display inflammation and does not usually require antimicrobial therapy. We agree that potential overdiagnosis of acute otitis media is a common problem and must be addressed for each child. Even among children with clinically diagnosed acute otitis media, however, many (80% in this trial) will experience resolution of the problem without specific antimicrobial therapy.

Both Pless and Kondzielewski found the double negatives in our interpretation baffling. We plead guilty to this charge, but note that equivalence and noninferiority trials are notoriously difficult to report. In our trial, for all children aged 6 months to 5 years, placebo was statistically inferior to amoxicillin. The CONSORT Group is presently working on a CONSORT extension for noninferiority trials.

## Footnotes

*   *Competing interests:* None declared.

## References

1.  1. Le Saux N, Gaboury I, Baird M, Klassen TP, MacCormick J, Blanchard C, et al. A randomized, double-blind, placebo-controlled noninferiority trial of amoxicillin for clinically diagnosed acute otitis media in children 6 months to 5 years of age. CMAJ 2005;172(3):335-41.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiY21haiI7czo1OiJyZXNpZCI7czo5OiIxNzIvMy8zMzUiO3M6NDoiYXRvbSI7czoyMjoiL2NtYWovMTczLzMvMjM2LjIuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

2.  2. Gomberg-Maitland M, Frison L, Halperin JL. Active-control clinical trials to establish equivalence or noninferiority: methodological and statistical concepts linked to quality. Am Heart J 2003;146(3):398-403.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1016/S0002-8703(03)00324-7&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=12947355&link_type=MED&atom=%2Fcmaj%2F173%2F3%2F236.2.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000185352800006&link_type=ISI) 

3.  3. Chen G, Wang YC, Chi GY. Hypotheses and type I error in active-control noninferiority trials. J Biopharm Stat 2004;14(2):301-13.
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=15206528&link_type=MED&atom=%2Fcmaj%2F173%2F3%2F236.2.atom)