ABSTRACT:The objective of this randomized clinical trial was to evaluate the clinical performance up to 18 months of restorations placed using ethanol-wet bonding technique (EWBT) compared with the three-step etch-and-rinse (TSER) and one-step self-etching (OSSE) approaches. Ninety-three non-carious cervical lesions (31 for each group) were restored by one experienced operator in 17 patients under relatively dry conditions using gingival retraction cord, cotton rolls and saliva ejector. Each adhesive system was randomly allocated to one of randomized cervical lesions until the three groups were present in the same subject in equal amounts. The restorations were evaluated at baseline, 6, 12 and 18 months by two blinded and calibrated examiners using the modified US Public Health Service guidelines (USPHS) for the following outcomes: retention (kappa= 1.00), staining and marginal adaptation (kappa=0.81) and analyzed by Fisher's exact and Kruskal-Wallis tests, respectively. No significant differences were observed among groups after 18 months for any of the assessed criteria (p>0.05). The intra-group analysis performed by Cochran's test (for retention) and Wilcoxon test (for marginal adaptation/staining) revealed significant differences between the time intervals baseline/18 months in marginal adaptation (p= 0.0117) and retention (p= 0.0101) for OSSE and in marginal staining for TSER (0.0051) and EWBT (p= 0.0277) groups. The survival analysis for retention criteria and the overall clinical success were performed using a log-rank test and did not show significant differences among groups (p> 0.05). All three adhesives protocols presented similar clinical performance up to 18 months.