This in vitro study compared the microleakage of Class V resin composite restorations at bevelled enamel/composite and dentin/composite interfaces following Er:YAG laser (pre-treatment modalities: laser-etching and/or acid-etching) or conventional preparation and acid-etch, in association with two resin composite formulations and their three-step adhesive system. Class V cavities with conventional bevel produced on the lingual and buccal surfaces of eighty extracted caries- and restoration-free human teeth, were assigned to eight groups: cavities were or Er:YAG-lased and acid-etched (groups 1 and 5); or Er:YAG-lased, laser-etched and acid-etched (groups 2 and 6); or Er:YAG-lased and only laser-etched (groups 3 and 7); or cut by dental drill at high-speed and acid-etched (groups 4 and 8). The specimens were restored with Optibond FL+Herculite XRV (groups 1, 2, 3 and 4) or with Scotchbond MP+Z 100 (groups 5, 6, 7 and 8), stored in distilled water at 37 degrees C for 24 h, thermocycled 1500 times between 5 and 55 degrees C, placed in a 2% aqueous solution of methylene blue for 24 h at 37 degrees C, embedded in resin and sectioned. Microleakage was assessed according to the depth of dye penetration along the restoration. There were statistically significant differences between occlusal and cervical regions for all groups (P<0.01) except for groups 3 and 7. Pair-wise comparison of groups showed that acid-etch is advocated when using resin composite in Er:YAG-lased Class V cavities; the seal at enamel margins in Er:YAG-lased and laser-etched cavities depended on the resin composite formulation and corresponding adhesive (P<0.05).