Purpose: To determine the marginal leakage of Class V restorations bonded with etch-and-rinse and self-etching adhesives applied after different dentin pretreatments over a maximum storage time of 24 months. Materials and Methods: Standardized mixed Class V cavities (5 mm x 3 mm and 2 mm deep) were cut on the buccal and lingual surfaces of 180 human molars. Two self-etching adhesive systems, Adper Prompt L-Pop (3M ESPE) and Clearfil SE Bond (Kuraray), and one etch-and-rinse bonding agent (One Step, Bisco) were applied as follows: 1. according to manufacturers' instructions; 2. after 37% H3PO4 etching for 15 s; 3. after 37% H3PO4 etching for 15 s and 5% NaOCl aq application for 2 min. Teeth were stored for 24 h, 6, 12, and 24 months in saline solution at 37°C before being stained in 0.5% solution of basic fuchsine. Dye penetration was scored on a 0 to 3 ordinal scale and analyzed with the Kruskal-Wallis H test (p 0.05), Mann-Whitney U-test (p 0.01), and Wilcoxon paired test (p 0.05). Results: Significant differences exist after using the tested adhesives at dentin and enamel margins. Adhesive type and substrate pretreatment had a significant effect on the long-term sealing of Class V restorations, and aging increased leakage overtime. The extent of leakage at the enamel margins was lower than that at dentin margins. One Step recorded the best results after 24 months. Conclusion: Optimal adhesion of restorative materials to enamel and dentin is hampered by a reduction in marginal seal over time. Alternative dentinal treatments (etching or collagen removal) might increase bonding efficacy, depending on the adhesive system used. Palabras clave: resin-infiltrated dentin, marginal integrity, microleakage, Class V restoration, bond longevity, substrate modification