To evaluate the effects of removing dentin collagen exposed by acid etching on the microleakage of bonded restorations with and without flowable composite application and submitted to thermocycling and long-term water storage. Standardized Class V cavities were prepared in 180 bovine incisors. They were randomly assigned to three groups according to the adhesive technique used. Conventional group (C): Single Bond was applied according to the manufacturer's instructions and Z250 composite was placed. Hypochlorite group (H): After acid etching a 10% NaOCl aqueous solution was applied for 1 min, then Single Bond and Z250 were applied. Hypochlorite and Flowable group (HF): Similar to group H, but following adhesive application, a thin layer of Natural Flow flowable composite was applied before the Z250. Each group was divided into three subgroups (0, 6, 12), which remained immersed in distilled water for 24 h or 6 or 12 months and underwent 500, 1500 or 2500 thermal cycles, respectively. At the end of each storage time, the specimes were stained with silver nitrate, decalcified, immersed in methyl salicylate for clearing and observed under a stereomicroscope to determine microleakage (scores 0 to 4). The data were analyzed using the Kruskal-Wallis and the Multiple Comparison Tests (α = 0.05). After 12 months, every group showed significant increases in microleakage. There was no significant difference between Groups H and HF for the three different periods of time, but they showed statistically less microleakage than Group C. The removal of dentinal collagen reduced the marginal microleakage when compared to the conventional technique. The use of the flowable composite did not produce significant effects. No technique was completely effective in preventing microleakage. Palabras clave: adhesion, microleakage, dentin, deproteinization, sodium hypochlorite