Statistics from Altmetric.com
A mother brings her 2 year old daughter to your clinic. She is concerned about her speech which she feels is poor for her age. Her daughter has failed two consecutive audiograms. On otoscopy she has signs of bilateral otitis media with effusion (OME) which you have confirmed on a previous occasion two months ago. Should you refer for insertion of grommets?
Structured clinical question
In preschool children with OME (glue ear) [patient] does the insertion of ventilation tubes (grommets) [intervention] as opposed to watchful waiting [comparison] have an effect on language development [outcome]?
Search strategy and outcome
Secondary searches: Cochrane—one relevant protocol (planned systematic review).4
Medline 1966 to November 2001, using the OVID interface: “otitis media with effusion”, “glue ear” , “otitis media” AND “middle ear ventilation” , “grommets” AND filter “clinical trial” , “controlled clinical trial” , “metanalysis” , “randomized clinical trial “ LIMIT to “English language”.
Search results—127 articles, three relevant. See table 3.
The use of language development as an outcome measure is problematic because of confounding factors, such as age, maternal education, and comorbid conditions. These studies have attempted to overcome this by randomisation, but still have problems with small study populations leaving the possibility of uneven allocation. The large numbers of the watchful waiting group being treated despite the study protocol may bias the results, reducing the apparent effect of ventilation tubes. These problems need to be taken into consideration when interpreting the studies.
In the UK, screening is not routinely carried out for OME, so the patient population in two of the studies is not similar to the UK population. The inclusion of “asymptomatic” children is likely to reduce the effectiveness of ventilation tubes, should any truly exist. The large unintentional crossover in some studies highlights the major problem in that there are no clear indications for the insertion of grommets.
Taken as a whole, there does seem to be some improvement in language skills a few months after the treatment. It is not evident that these effects persist long term and the differences seem to diminish with time, as one would expect with the natural history of the condition.
As grommet insertion is the most common elective operation in preschool children, concerns both about risks of the anaesthesia and cost-benefit analysis of the procedure need to be addressed. The results of a further randomised controlled trial are awaiting publication.5
▸ Clinical bottom line
In children with OME and language delay, there is no good evidence to suggest that insertion of ventilation tubes will improve language development.
A more important factor is mother’s level of education, which has been shown to have a greater effect on language acquisition in young children.
Table 3 Grommets in OME
Citation Study group Study type (level of evidence) Outcome Key results Comments
Rach et al (1991) n=52 Age: 2 year olds Children from a larger cohort were screened for OME with tympanograms Randomised to treatment Randomised controlled trial (level 2b) Improvement in language scores at 6 months Improvement in scores in the treatment group, but large overlapping CI Follow up 100%; small numbers in trial and follow up time not sufficient Balanced randomisation Testers not blinded Maw et al (1999) n=182 Age: 2–3 year olds Confirmed to have OME and hearing loss with tympanograms; only those with problems in speech, learning, or behaviour were included Randomised to treatment Randomised controlled trial (level 2b) Improvement in language scores at 9 and 18 months Improvement in treated group at 9 months Mean difference between groups was not significant 0.31 (-0.03 to 0.66) 95% CI At 18 months smaller differences which were not significant Follow up 83% Large drop out in trial makes results difficult to interpret; by 18 months 85% of watchful waiting group had grommets inserted Data analysed by intention to treat Rovers et al (2000) n=187 Age: 16–24 months From a larger cohort that were screened for hearing loss and OME Randomised to treatment Randomised controlled trial (level 2b) Improvement of language scores at 6 and 12 months No difference between treatment and watchful waiting groups Follow up 79% Balanced allocation for groups but groups were different for confounding factors; not clear if testers were blinded