In vitro microleakage associated with posterior composite restorations used with different base/bonding system combinations.

The study aimed to determine the base/restoration combination which allowed the least microleakage in Class II cavities. One hundred premolar teeth were prepared with a standardized, minimal two-surface cavity and randomly ascribed to ten equal groups. The cavities were restored by use of one of ten different material/technique combinations. Restored teeth were sealed to within 1 mm of the restoration margins with nail varnish prior to the immersion of the crowns in 5% (buffered) eosin for 48 h. The teeth were then hemisected longitudinally, in a mesio-distal plane, through the midline of the restoration by means of a diamond disc with water coolant, and one half of each tooth was then photographed. The transparencies were examined by image analysis programmed to facilitate calculation of the percentage of the tooth/restoration interface and of the crown dentin exhibiting dye penetration. Light- and chemically-cured glass-ionomer-based restorations showed similar amounts of leakage when expressed as a percentage area of crown dentin. This was significantly (p less than 0.05) less than that shown by cavities both etched and bonded. Cavities restored without enamel etching or involving the use of a bonding system exhibited significantly (p less than 0.05) greatest leakage.