The marginal seal of composite resin restorations was compared in vitro between the preventive class II restoration (Surmont,1990) and the conventional class II restoration, and also between the 15-second etching using phosphoric acid and maleic acid. The shape of the proximal box followed that for the traditional class II amalgam restoration for either preparation and the shape and size were approximately standardized. For the preventive restoration, extension through the occlusal groove was minimized and no dove-tail retention form was applied to use the sealant for sealing the rest of the pits and fissures. All of the preparations and restorations were performed on extracted human premolars, using Scotchbond^TM Multi-Purpose system and Z-100 (3 M), except for the comparative use of phosphoric acid etchant (Scotchbond^TM Etchant). Repeated compressive stresses were applied on the isthmus area on the composite resin restoration for 10,000 cycles, or thermal stresses of 40°C and 60°C were applied alternately for 100 cycles. The dye penetration test was then performed to evaluate the microleakage at the cervical, proximal, and occlusal margins.
The results showed that no statistical difference in microleakage was found between the preventive and conventional restorations in both the maleic acid and phosphoric acid groups. However, the phosphoric acid group generally demonstrated a better marginal seal than the maleic acid group except for the cervical margins. Although no clear difference was found at the cervical margins among experimental groups, most cases prevented microleakage into the dentin.