Purpose: To compare a self-etch and a two-step etch-and-rinse adhesive in terms of internal and marginal composite-tooth bond failure separately on enamel and dentin/cement at 36–48 months after restoration placement using optical coherence tomography (OCT).
Materials and Methods: Twenty-seven patients with two or three class V composite restorations of noncarious cervical lesions 36–48 months after placement were included. The one-step self-etch adhesive Futurabond M ([Voco] group SE, n = 25) and the two-step etch-and-rinse adhesive Solobond M ([Voco] group ER, n = 20) combined with the nanohybrid composite Amaris (Voco) were evaluated. The four-step etch-and-rinse adhesive Syntac classic combined with Tetric EvoCeram (Ivoclar Vivadent) served as the control (n = 18). Spectral-domain OCT (SD-OCT, 1310-nm center wavelength) was applied. Marginal gaps and internal interfacial adhesive defects were quantified in cross-sectional OCT images. Groups were statistically compared using the Friedman/Wilcoxon test (α = 0.05).
Results: In enamel, nonsignificantly different percentages of marginal gap formation and internal interfacial adhesive defects were found between the groups (pi ≥ 0.258). In dentin/cement, SE showed significantly less marginal gap formation compared to ER (p < 0.001) and control (p = 0.001), and at the internal dentin-composite interface less adhesive defects were found compared to ER (p < 0.001) and control (p = 0.003).
Conclusion: The self-etch adhesive used in the current study appears recommendable for restoration of noncarious cervical lesions with composite. Keywords: class V composite restoration, self-etching adhesive, etch-and-rinse adhesive, OCT, internal interfacial gap formation, quantitative margin analysis.