Comparison of three methods of critical evaluation of microleakage along restorative interfaces.

This in vitro study compared microleakage along the dentin-restorative interface by use of a conventional single-surface and two stereoscopical protocols. The extent of microleakage in class V cylindric cavities that were restored with Scotchbond 2/P-50 was assessed after thermocycling and silver staining were performed. Specimens were randomized into three groups of 10 teeth for the three different protocols to be investigated. The categories were group 1, a single longitudinal section through the midline of the restoration, group 2, a multiple-surface protocol based on computer reconstruction of digitized images from multiple serial abrasions, and group 3, a clearing protocol based on immersion of demineralized specimens in methyl salicylate. The severity of microleakage for all groups was recorded on an ordinal scale and based on linear leakage length. The median leakage score for groups 1 through 3 was 2, 4, and 4, respectively. When comparisons were performed with nonparametric statistical analysis, both the clearing and the multiple-surface protocols recorded significantly more severe leakage than those detected from a conventional, single longitudinal, midline sectioning technique, (x2 = 8.595, p < 0.02). Furthermore microleakage patterns generated from the two stereoscopical protocols revealed extensive microleakage that advanced unilaterally or bilaterally through the dentin-restorative interface along the embrasure regions into the axial wall of the restorations. No significant difference was noted between the two stereoscopical protocols (Q = 0.260, p > 0.05), which suggested that the clearing protocol, which is less labor-intensive than the multiple-surface protocol, may be recommended for future comparative microleakage studies that involve the dentin-restorative interface.