General Practice

Commentary: Reduced confusion over ear effusion?

BMJ 1997; 314 doi: (Published 01 February 1997) Cite this as: BMJ 1997;314:354
  1. M P Haggard, directora
  1. a MRC Institution of Hearing Research, Nottingham NG7 2RD


    Otitis media with effusion (glue ear) is resource consuming and confusing because of its fluctuating course, eventual age related remission, and variability both in presentation and consequences. The cumulative incidence is extremely high, and the point prevalence distributions for severity (degree of hearing loss) and for persistence (time spent with otitis media with effusion) have no discontinuities; only a statistical cut off point is therefore feasible between a case and a non-case.1 A treatable instance cannot be defined by an examination or test on one particular day, which merely confirms diagnosis. This importance of total histories requires two medical responses: (a) much reliance on parental reporting of detail and extent, accompanied by vigilant awareness that some parents underreport, and (b) selective referral on the basis of risk factors for more objective tests to offset such underreporting.

    The work of Hogan et al updates our understanding of otitis media with effusion in several ways. It offers accurate descriptive data on early histories with frequent time sampling, summarised by parameters of the long established Markov class of model. Although this class needs to be tested against deterministic alternatives–for example, chaotic oscillatory models–for goodness of fit, its parameters probably offer useful descriptors for children, risk factors, and treatments. The finding that apparent persistence is due chiefly to recurrence further questions the value of a single test on an arbitrary day, and hence of any universal screen for otitis media with effusion. It also supports an important recent revision of our view of the cause of the disease. The new view invokes a continuing low grade infection maintained by the effusion that can flare up, leading back to a subacute inflammatory state, without evident exposure to new infection immediately beforehand. The effusion aspirated at myringotomy often does not culture bacteria by conventional methods, but it is not sterile as formerly thought. Many such culture negative samples contain bacterial DNA, suggesting viable bacteria.2 Meta-analysis shows that, although antibiotics are effective in acute otitis media, the magnitude and duration of their effect in otitis media with effusion do not currently offer a satisfactory treatment policy.3 Drug resistant bacteria are bred from overprescription of antibiotics in the community, largely for otitis media,4 though resistant strains in themselves are unlikely to explain the allegedly increased prevalence of otitis media with effusion in recent decades. The evidence for partial efficacy of antibiotics in this disease suggests two future possibilities: (a) more scope for new medical treatments and (b) a large pragmatic effectiveness trial to define prescription constraints for medium length courses of antibiotics as temporising management for otitis media with effusion, thereby potentially covering one part of the continuum of high parental demand. The Royal Society of Medicine Encyclopaedia of Child Health may be pardoned for telling parents at this point that otitis media with effusion is not primarily caused by infection,5 as more epidemiology, management trials, and dissemination of information are still required to put this revised account of the disease fully to work.

    The guerilla war fought in the pharmacokinetic backwaters of the middle ear mucosa between bacteria synergised by viruses and the immature immune system, sometimes aided by modest local antibiotic concentrations, is too easily ignored. Its slowness and precarious complexity make the longstanding confusion of professionals and parents less surprising.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    View Abstract