Comparative in vivo and in vitro investigation of interfacial bond variability.

This comparative in vivo/in vitro study investigated the interfacial adaptation between dentin and composite resins in totally bonded adhesive restorations placed under clinical and laboratory conditions in the same tooth. Cavities were prepared in buccal or lingual surfaces of 47 third-molar teeth scheduled for extraction and randomly assigned for treatment with the following bonding systems/restorative materials: Clearfil Liner Bond 2/Clearfil AP-X (Group I, n = 10), Resulcin AquaPrime + Monobond/MFR Merz (Group II, n = 9), Prime&Bond 2.1/Dyract AP (Group III, n = 9), Scotchbond Multi-Purpose/Z-100 (Group IV, n = 10), and Ecusit Primer AB-Mono/Ecusit (Group V, n = 9). In Group V, a thin layer of a low-viscous compomer (Primaflow) was interposed between the dentin and the composite resin during restoration placement (according to the manufacturer's directions). After extraction a second filling was placed in each tooth with identical materials in the same manner as in vivo. The restoration-dentin interface was evaluated in longitudinally cut sections of the specimens by SEM-analysis, and the frequency of gap formation between restoration and dentin was calculated. Median percentages for the in vivo/in vitro frequency of interfacial gap formation were 29.2%/13.9% in Group I, 33.3%/20.0% in Group II, 40%/5.3% in Group III, 53.9%/30.4% in Group IV and 13.8%/0% in Group V. Comparison of gap formation frequency between fillings placed in vivo and in vitro revealed significant differences (p < 0.05) in Groups II, III and IV. However, in Groups I and V the internal adaptation was not significantly different between in vivo and in vitro applied restorations. A significant (p < 0.01) correlation (Spearman-Rho rank correlation coefficient) was found for the corresponding in vivo and in vitro fillings placed in the individual teeth with regard to interfacial gap formation. It was concluded that achievement of a completely gap-free internal adaptation between restorative material and dentin in totally bonded composite resin restorations is difficult to predict under in vivo as well as in vitro conditions.