Parental smoking, middle ear disease and adenotonsillectomy in children

vestigated the association of exposure to environmental tobacco smoke and diseases of the Background – A systematic quantitative review was conducted of evidence relating ear, nose and throat (ENT), and the evidence has been the subject of periodic narrative reparental smoking to acute otitis media, recurrent otitis media, middle ear effuviews. Only two studies have attempted a quantitative meta-analysis. 6 In one, based on sion, and adenoidectomy and/or tonsillectomy. papers published up to 1992 inclusive, the authors did not distinguish clearly between Methods – Forty five relevant publications were identified after consideration of 692 studies of different types of ENT disease. The other, based on literature published up to articles selected by electronic search of the Embase and Medline databases using 1994, reviewed parental smoking briefly along with other risk factors for acute otitis media. keywords relevant to passive smoking in children. The search was completed in This paper systematically reviews the evidence relating parental smoking to acute otitis April 1997 and identified 13 studies of acute otitis media, nine of recurrent otitis media, recurrent otitis media, “glue ear” (otitis media with effusion), and ENT surgery in chilmedia, five of middle ear effusion, nine of glue ear surgery, and four of adenodren. Each of these outcomes is considered separately, including a quantitative metatonsillectomy. A quantitative meta-analysis was possible for all outcomes except analysis where appropriate. acute otitis media, using random effects modelling where appropriate to pool odds ratios from each study. Methods Results – Evidence for middle ear disease Published papers, letters, and review articles is remarkably consistent, with pooled odds were selected by an electronic search of the ratios if either parent smoked of 1.48 (95% Embase and Medline databases using the CI 1.08 to 2.04) for recurrent otitis media, search strategy described in detail elsewhere. 1.38 (1.23 to 1.55) for middle ear effusion, Briefly, all passive smoking references were and 1.21 (0.95 to 1.53) for outpatient or selected by the MESH heading tobacco smoke inpatient referral for glue ear. Odds ratios pollution and/or textword combinations ({passfor acute otitis media are in the range ive, second-hand, second hand, involuntary, par1.0 to 1.6. No single study simultaneously ent∗, maternal, mother∗, paternal, father∗ or addresses selection bias, information bias household} and {smok∗, tobacco∗ or cigarette∗}). and confounding, but where these have Papers were then restricted to children by relbeen investigated or excluded in the design evant textwords or by the age group as specified or analysis, the associations with parental in the title or abstract. This search, completed smoking persist virtually unchanged. in April 1997, yielded 3625 references of which Large French and British studies are in1593 contained keywords relevant to resconsistent with regard to the association piratory or allergic disease. These 1593 abof parental smoking and tonsillectomy. stracts were reviewed and 99 papers relevant Conclusions – There is likely to be a causal to ENT disease were selected by the textwords relationship between parental smoking tympanom∗, otitis, middle ear, glue ear or tonsil∗. and both acute and chronic middle ear Among these 99 papers, 37 publications predisease in children. sented quantitative information relevant to this Department of (Thorax 1998;53:50–56) Public Health review and a further six were identified by Sciences, citations in previous overviews or individual