No research evidence to support the prophylactic removal of pathology-free impacted third molars

Abstract

Objective To summarise existing evidence on prophylactic removal of impacted wisdom teeth, in terms of the incidence of surgical complications associated with prophylactic removal and the morbidity associated with retention.Data sources Update of a previous review with additional searches of MEDLINE (1984–1999), EMBASE (1984–1999), Science Citation Index, Cochrane Controlled Trials Register, The National Research Register, The Database of Abstracts of Reviews of Effectiveness, and the NHS Economic Evaluation Database plus additional paper and web-based sources.Study selection Reviews or RCTs that considered surgical removal of unerupted or impacted third molars, either as prophylaxis or because of pathology reporting outcomes including pathology and/or symptoms associated with unerupted or impacted third molars or surgical outcomes were eligible for inclusion.Results We identified 4682 references, of which 28 met the inclusion criteria. Two RCTs, 22 literature reviews and four decision analyses. Results from both RCTs suggested that prophylactic removal of third molars is unjustified (one trial provides preliminary data only). Methodological quality of the reviews was generally poor. Five reviews suggested only a weak association between retention of third molars and anterior crowding. Five reviews concluded that the prophylactic removal of third molars was unjustified. Nine reviews had no clear conclusion. One review (of poorer methodological quality) suggested that prophylactic removal of third molars is appropriate. The decision analyses indicated that retention of third molars was cost saving and more cost-effective compared with prophylactic removal of all impacted third molars. There are currently two ongoing RCTs, one in Denmark the other in the USA.Conclusion There is no reliable research evidence to support the prophylactic removal of pathology-free impacted third molars. Well-designed RCTs with long-term follow-up are required in order to confirm this.