Otitis media with effusion (OME) is the most prevalent form of middle ear disease in young children. It is defined as the presence of fluid behind the tympanic membrane without the symptoms or signs of acute otitis media (AOM).1 It is usually associated with some hearing loss. OME with persistent hearing loss may contribute to delays in speech and language development.2 Recommended interventions include the use of antibiotics and the insertion of ventilation tubes (grommets).2-4

There have been concerns about the overuse of grommet surgery. Substantial variation in rates of this procedure have been documented.5,6 Clinical practice guidelines have recommended that grommets are an option for children who have bilateral OME associated with a hearing loss of > 20 decibels for at least 3 months.2-4 The intervention is most likely to benefit younger children (during the most critical phase of language development) and those with most hearing loss.

This trial was part of the largest otitis media cohort study ever conducted.7 The study was generally well designed and well reported.8,9 A total of 6350 healthy infants were enrolled in their first 2 months of life and were evaluated monthly for otitis media; 588 of these children had persistent or very frequent OME. They were regarded as typical of children who might receive surgery in the United States. The aim of the study was to determine the benefits of early referral for surgery compared with delayed referral (where “watchful waiting” continued for an additional 6 months).

Random assignment was made by designated non-clinical staff using separate, computer-generated lists of random numbers. Children were stratified according to site, age (in 6-month categories), and whether the eligibility criteria were met on the basis of bilateral or unilateral effusion. Assignment within each of the predetermined strata occurred in permuted blocks of four. This ensured that the allocation was balanced after every four new children (in a stratum) were randomly allocated.

Children underwent developmental assessment as soon as possible after their third birthday (and always within 2 months). Standardised assessment tools were used. Assessors were unaware of the child’s medical history, health insurance status, and mother’s level of education. The investigators proposed that a difference of 0.33 standard deviations between groups on any outcome measure could be clinically important. Follow-up of participants over a prolonged period was reasonably good (95% and 92% of the early- and late-treatment groups, respectively). All analyses were based on the intention-to-treat principle (although children who were not assessed could not be included in the analysis).

Any weaknesses in the methods and the quality of reporting were relatively minor. Ideally, the authors should have also described (i) how random allocation was concealed from the investigators, (ii) how block size was concealed from investigators (so they couldn’t guess which intervention would be allocated next), (iii) the adequacy of blinding, (iv) the primary outcome for the study, and (v) an assessment of adverse outcomes. The choice of a 0.33-standard- deviation difference between the groups in any of the assessments as the minimal clinically important difference was controversial. While this approach should be able to identify reasonably small statistical differences attributable to the intervention, the clinical importance of such a difference is not easily understood. In the end, because there were no statistically significant differences in any of the outcome measures, this issue did not arise.

This is the largest randomised trial to evaluate ventilation tubes in children with persistent OME. Previous studies had shown that surgery improved hearing by around 12 dB at 6 months and 6 dB at 12 months.5 Most of these studies involved older children and did not include an assessment of speech and language (which is generally regarded as the most important outcome).

The results of this study are consistent with those of previous studies in demonstrating that grommet surgery will substantially reduce the amount of time that a child has OME, and modestly improve hearing. However, by 3 years of age, early referral for surgery had no beneficial effect on development or behaviour. The consistent lack of effect for a range of outcome measurements was striking. These findings were not changed by the subsequent subgroup analyses.10 The results of this study are unlikely to be explained by a biased estimate of effect or by chance. Similar results have also been documented in other recent well designed studies in different populations.11-13