OBJECTIVES: To evaluate the clinical performance of a universal adhesive system containing 0.1 % copper nanoparticles (CuNp) in non-carious cervical lesions (NCCLs) after 6 years. MATERIAL AND METHODS: Two hundred and sixteen restorations (n = 54) were randomly placed in 36 participants with at least four NCCLs. Groups included: ER-Cu (adhesive in etch-and-rinse strategy with CuNp), ER-Ct (etch-and-rinse without CuNp), SE-Cu (self-etch with CuNp), and SE-Ct (self-etch without CuNp). Restorations were evaluated at baseline and after 1, 3, 4, and 6 years using the FDI and USPHS criteria. PRIMARY OUTCOME: fracture of material and retention. Examiners and patients were blinded to group assignment. RESULTS: After 6 years, 104 restorations were analyzed (ER-Cu 24, ER-Ct 29, SE-Cu 27, SE-Ct 24). Fourteen restorations loss retention (ER-Cu 6, ER-Ct 1, SE-Cu 2, SE-Ct 7). Retention rates (95 % confidence interval) were 79.3 % (61.6 - 90.1) for ER-Cu
97.0 % (83.3 - 99.4) for ER-Ct
93.1 % (78.1 - 98.1) for SE-Cu
and 76.7 % (59.1 - 88.2) for SE-Ct, with statistical differences between SE-Cu vs. ER-Ct, and SE-Ct vs. ER-Ct (p <
0.05). Thirty-one restorations (ER-Cu 6, ER-Ct 5, SE-Cu 9, SE-Ct 11) showed minor marginal staining, and forty-one restorations (ER-Cu 8, ER-Ct 10, SE-Cu 11, SE-Ct 12) presented minimal marginal adaptation defects (p >
0.05). No significant differences were found for other secondary parameters (p >
0.05). CONCLUSION: This is the first 6-year clinical trial showing the long-term behavior of a universal adhesive system in NCCLs. The addition of CuNp does not affect the clinical performance of the universal adhesive.