Dentin bond strength and marginal adaptation: direct composite resins vs ceramic inlays.

The aim of this in vitro study was to compare the dentin bond strength and marginal adaptation of directly and indirectly inserted restorations. A conically modified push-out test was designed to consider polymerization shrinkage and facilitate inlay placement. A total of 260 cavities were prepared into disks of freshly extracted human third molars and filled with direct composite resins or with adhesively luted ceramic inlays. Dentin adhesives of the third--(with self-etching primer: ART Bond, Syntac Classic), fourth--(with total etching: Scotchbond Multi-Purpose Plus), and fifth-generation (one-bottle adhesives: Syntac Single Component, Prime & Bond 2.1) were used in combination with one hybrid composite (Tetric) or luting composite (Variolink Low). Control groups did not use an adhesive. Polymerization of the bonding agent was carried out prior to insertion of the filling/inlay or afterwards simultaneously with the composite/luting composite. The thickness of the adhesive layer and luting composite was recorded, and after 7 days of storage and 24 hours of thermocycling (1150 cycles) replicas were made and extrusion testing performed. Fracture modes were determined and replicas were examined regarding marginal adaptation using SEM (X200 magnification). Precuring of the bonding resin increased dentin bond strength independent of the material combination or insertion mode (P < 0.05). In general, third- and fourth-generation dentin adhesives produced better results in bond strength and marginal adaptation than one-bottle systems (P < 0.05). In the third generation, ART Bond achieved significantly higher push-out values than Syntac (P < 0.05), but no better marginal adaptation. Cohesive fractures within the dentin were only observed in the inlay groups with precured resin. Precuring of the bonding resin is an important factor for both direct and indirect restorations. Nevertheless, precuring of the bonding resin prior to insertion of adhesive inlays cannot be recommended clinically, because the 120-micron luting spaces were too large. In simulated cavities, direct composite fillings with precuring achieved bond strengths similar to inlays without precuring. One-bottle adhesive systems performed poorly compared with multi-step adhesives of the third and fourth generation.