The effect of tooth preparation on microleakage behavior

Many factors contribute to the microleakage of a restoration. One of the more important is the method of cavity preparation. This study compared the microleakage behavior of composite restorations placed in cavities prepared by different techniques. It also compared and correlated the microleakage data produced by an electrochemical vs a staining technique. Class V cavities were prepared in 48 premolars by four techniques: (1) tungsten carbide bur in a high-speed handpiece followed by acid etching; (2) air abrasion (27 microns Al2O3) followed by acid etching; (3) air abrasion (50 microns Al2O3) and (4) air abrasion (27 microns Al2O3), with n = 12 in each group. All teeth were restored with Prime and Bond 2.1 and Tetric Flow, then thermocycled between 5 degrees and 55 degrees C for 5000 cycles with a one minute dwell at each temperature. After thermocycling, a PVC-covered Cu wire was inserted apically into the pulp chamber of each tooth and sealed into position. Leakage was continuously followed by a conductimetric method for 75 days. The teeth then were immersed in 50% AgNO3 for two hours, rinsed in distilled water for 60 seconds, then placed in a rapid photographic developer solution for two hours, followed by rinsing and sectioning for microscopic examination. Electrochemical data were examined by ANOVA and Newman-Keuls multiple comparison tests, while Kruskal-Wallis and Rank Sum Difference tests were used on the staining evaluations. Spearman's rho test was used to correlate the two test techniques. Electrochemical data for cavities prepared with a bur or air abrasion followed by acid etching prior to restoration showed significantly less (p < or = 0.05) microleakage (mean leakage currents of 1.89 & 1.57 microA, respectively) than teeth prepared with air abrasion alone (mean leakage currents of 3.60 & 3.40 microA, respectively). Rank sum AgNO3 staining data (196 & 242 vs 371 & 368) supported these findings. The correlation between the electrochemical and staining data was significant (p < or = 0.05) for all four groups of test specimens.