OBJECTIVE The successful addition of new restorative materials to an existing restoration may be the most conservative course of treatment. Repairing amalgam restorations with resin materials remains a viable clinical alternative to amalgam replacement. This in vitro study evaluated the effect of different adhesive systems and surface treatments on the integrity of amalgam-resin and resin-tooth interface after partial removal of pre-existing amalgam. METHODS Fifty defect-free human molars were restored with amalgam occlusally. The teeth were thermocycled (1,000x) between 5 degrees C and 55 degrees C, with a dwell time of 30 seconds. The mesial and distal parts of the amalgam fillings were removed, leaving only the middle part of amalgam. One side of the cavity was finished with a coarse diamond bur, while the other part of the amalgam was finished with a fine diamond bur. The samples were then randomly divided into five groups (n = 10/group) and received the following adhesive systems: Group 1: All Bond 3 (BISCO, Inc); Group 2: Clearfil SE Bond+Alloy Primer (Kuraray); Group 3: Kuraray DC Bond (Kuraray); Group 4: Xeno V (Dentsply); Group 5: XP Bond (Dentsply). All the cavities were restored with resin composite (TPH Spectrum, Dentsply). All the materials were used according to the manufacturer's directions. The specimens were re-thermocycled (1,000x), sealed with nail varnish, stained with 0.5% basic fuchsin for 24 hours, sectioned mesiodistally and photographed digitally. The extent of dye penetration on the tooth-sealant interface was measured by image analysis software (ImageJ, Scion Image, Frederick, Maryland, USA) for both coarse-finished and fine-finished surfaces at the resin-tooth and resin-amalgam interface. The data were analyzed statistically with one-way ANOVA and post hoc Tukey tests (alpha = 0.05). RESULTS All Bond 3 and XP bond (etch & rinse) produced the best results at each section. All the materials exhibited more microleakage at the amalgam interface than the tooth interface. Surface finishing with different burs did not statistically affect microleakage. CONCLUSIONS In terms of microleakage reduction, etch & rinse adhesives may be preferred over self-etch adhesives for amalgam repair.