Purpose: To evaluate the effect of direction of first cure and the presence or absence of a facial margin bevel on the adaptation of facial margins of Class III resin-based composite (RBC) restorations.
Materials and methods: Custom polyvinyl siloxane (PVS) light shields were fabricated for both the mesial and distal surfaces of 20 extracted incisors. Class III cavities were prepared in the mesial and distal surfaces of the incisors. On the facial, lingual, and gingival margins of 20 of the preparations a 0.5 mm wide enamel bevel was placed. The lingual and gingival, but not the facial, margins were beveled on the other 20 preparations. The PVS light shield was adapted to the tooth and trimmed to allow 1-2 mm of the facial margin to be exposed on 10 of the beveled and 10 of the non-beveled preparations. For the remaining 20 preparations the PVS light shield was trimmed to allow access of the curing light only from the lingual. Teeth were restored with Prisma TPH RBC. Using the appropriate shield, restorations in the facial first cure group were cured first from the facial and then from the lingual. Restorations in the lingual first cure group were cured only from the lingual. Restorations were finished flush with enamel margins and thermocycled following storage. Samples were exposed to a 50% solution of silver nitrate followed by light exposure. Teeth were sectioned and microleakage was evaluated at three levels (incisal, middle, and cervical). Four evaluators (blinded to direction of cure) independently scored microleakage using a categorical evaluation scale. Statistical analysis included non-parametric descriptive statistics, Cohen's kappa, chi-square analysis, and Kruskal-Wallis ANOVA.
Results: Samples prepared without a facial bevel revealed significantly more (P = 0.01) microleakage than preparations with a facial bevel. At the most incisal level, restorations cured first from the facial revealed significantly more (P < 0.001) microleakage than restorations cured first from the lingual. At the middle and cervical levels, restorations cured first from the facial revealed significantly less (P < 0.001) microleakage than restorations cured first from the lingual. When a facial bevel was present, samples cured first from the facial showed significantly less (P < 0.001) microleakage than those cured only from the lingual.