BACKGROUND Dental caries (tooth decay) is one of the commonest diseases which afflicts mankind, and has been estimated to affect up to 80% of people in high-income countries. Caries adversely affects and progressively destroys the tissues of the tooth, including the dental pulp (nerve), leaving teeth unsightly, weakened and with impaired function. The treatment of lesions of dental caries, which are progressing through dentine and have caused the formation of a cavity, involves the provision of dental restorations (fillings). OBJECTIVE To assess the effects of adhesive bonding on the in-service performance and longevity of restorations of dental amalgam. METHODS Databases searched July 2009: the Cochrane Oral Health Group's Trials Register; CENTRAL (The Cochrane Library 2009, Issue 3); MEDLINE (1950 to July 2009); and EMBASE (1980 to July 2009). METHODS Randomised controlled trials comparing adhesively bonded versus traditional non-bonded amalgam restorations in conventional preparations utilising deliberate retention, in adults with permanent molar and premolar teeth suitable for Class I and II amalgam restorations only. METHODS Two review authors independently screened papers, extracted trial details and assessed the risk of bias in the included study. RESULTS One trial with 31 patients who received 113 restorations was included. At 2 years only 3 out of 53 restorations in the non-bonded group were lost, which was attributed to a lack of retention, and 55 of 60 bonded restorations survived with five unaccounted for at follow-up. Post-insertion sensitivity was not significantly different (P > 0.05) at baseline or 2-year follow-up. No fractures of tooth tissue were reported and there was no significant difference between the groups or matched pairs of restorations in their marginal adaptation (P > 0.05). CONCLUSIONS There is no evidence to either claim or refute a difference in survival between bonded and non-bonded amalgam restorations. This review only found one methodologically sound but somewhat under-reported trial. This trial did not find any significant difference in the in-service performance of moderately sized adhesively bonded amalgam restorations, in terms of their survival rate and marginal integrity, in comparison to non-bonded amalgam restorations over a 2-year period. In view of the lack of evidence on the additional benefit of adhesively bonding amalgam in comparison with non-bonded amalgam, it is important that clinicians are mindful of the additional costs that may be incurred.