Objectives: The aims of this study were to compare the local and systemic uptake of fluoride released from a compomer material (Dyract Ortho) and a resin-modified glass ionomer cement (Vitremer) with that of a conventional resin adhesive (Right-On) and to compare the cariostatic ability of each of the test materials with that of the resin control. Methods: Twenty six patients were randomly allocated to have a bracket bonded to a premolar on one side of the arch with one of the test materials and on the opposite side with the control material. Premolars destined for extraction as part of an orthodontic treatment plan were selected for bonding. A non-fluoride toothpaste was used by all participants for 4 weeks prior to bracket bonding and throughout the 4 week trial period. Fluoride release was measured in saliva, plaque and urine samples taken pre-bonding and 4 weeks post-bonding. Enamel demineralisation was assessed by scoring the buccal surface of each extracted tooth using a caries index. Results: Neither Vitremer nor Dyract Ortho altered salivary or urinary fluoride concentration significantly 4 weeks post-bonding but plaque fluoride concentration increased significantly around premolars bonded with Vitremer. The test materials as a combined group were associated with significantly less demineralisation than the control material but there was no significant difference in cariostatic ability detected between either Dyract Ortho or Vitremer when each group was compared separately with the control. Conclusions: Fluoride released from Dyract Ortho or Vitremer is likely to exert a local and not a systemic effect. In a 4-week clinical study, the cariostatic ability of the fluoride-releasing cements, as a combined group, was superior to that of the non-fluoride releasing control but there was no significant difference in cariostatic ability between the two test materials when each test group was compared separately with the control. (C) 1998 Elsevier Science Ltd. All rights reserved.