There is only limited knowledge of the factors which influence the outcome of otitis media with effusion in children in the long-term. This randomized controlled study assessed the therapeutic effect of adenoidectomy and adenotonsillectomy during a 5-year follow-up. Numerous pre-treatment independent variables concerning the child's upper and lower respiratory tract, atopic status and parental habits were assessed in relation to two dependent outcome measures. These were otoscopic clearance of effusion and no peak/peak tympanometric change. A total of 222 children was studied and reviewed annually for 5 years. Four of 43 independent variables were found to be repeatedly significant in relation to outcome: (a) whether or not adenoidectomy was performed; (b) age at operation; (c) history of earache prior to operation; and (d) parental smoking habits. The results provide further evidence of a beneficial effect of adenoid removal and the importance of the age at which surgery is advised. They also suggest the need to investigate further the relationship of superadded acute suppurative otitis media with otalgia and the outcome of chronic otitis media with effusion. Finally, avoidance of parental smoking will have a beneficial effect on children's middle ear disease.