To increase bond strength of brackets bonded to composite restorations, underlying composite restorations are usually roughened. This and subsequent bracket debonding might deteriorate smoothness of composite restoration, leading to clinical and aesthetic problems. Since this was not assessed before, this study was conducted.

Seventy intact human premolars were collected and filled with composite restorations. They were randomly divided into 3 groups: control (n=10), surface treatments before bracket bonding (bur, sandblasting, overall n=40), and surface roughening without bracket bonding (bur, sandblasting, overall n=20). The 40 specimens with bracket bonding steps were divided into two subgroups of with and without polishing after debonding of brackets (each n=10). After aging the composites in two steps and treating surfaces according to abovementioned protocols, their surface topographies were evaluated under stereomicroscopy (80×), scanning electron microscopy (using Zachrisson-Arthun index, at 500× and 1000× zooms), and by profilometry. Profilometry parameters and Zachrisson-Arthun index were statistically compared among different groups (α=0.05, 0.005).

According to Kruskal-Wallis test, all profilometry parameters or the Zachrisson and Arthun index were significantly different between groups involving bracket debonding and control (P≤0.001). Three-way ANOVA indicated that bracket debonding, polishing after debonding, and surface treatments before bracket bonding had a significant effect on Ra and Rz (P≤0.05). Tukey post hoc test showed that bur and sandblasting methods were not significantly different (P>0.05).

Bur-roughening or sandblasting of composite surface before bonding brackets can increase surface roughness, while polishing can reduce it back to control levels.