SUMMARYObjective: To investigate the effectiveness of alternative treatments for replacing defective amalgam restorations through a prospective longitudinal cohort clinical study. Methods: Fortyfive patients aged 21 through 77 (mean=56) years with 113 defective amalgam restorations, which were independently diagnosed during treatment planning, participated in the study. These patients were assigned to 5 treatment groups: repair (n=20), sealing of defective margins (n=23), refurbishing (n=23), replacement (n=23) and notreatment (n=24). The replacement and no-treatment groups served as comparison groups and received random assignment. Two clinicians examined the restorations (n=113) prior to and after the assigned treatment and at subsequent recalls, using a modified Ryge Criteria that included marginal adaptation, anatomy, contact, post-operative sensitivity and secondary caries. Results: At 1-and 2-year recalls, 79 (70%) and 74 (65%) restorations were examined. KruskalWallis Test showed significant differences for marginal adaptation and anatomic form for both 1-and 2-year recall exams (p<.05). The repair and replacement groups had significant differences when compared to the no-treatment group. Conclusions: Defective restorations that have a Bravo rating for clinical characteristics other than marginal integrity and anatomical form do not need to be immediately replaced.