This study investigated the influence of dentin conditioning and contamination on the marginal adaptation of Class II sandwich restorations. Large butt-joint MOD cavities with cervical margins located 1 mm below the CEJ were cut into 72 extracted human molars. Nine groups were filled using a total-bond technique with Z100 or a sandwich technique with either Vitremer or F2000 in combination with Z100. For all three material combinations three different pretreatments were compared: total etch, selective etch and dentin contamination with saliva and blood prior to primer/adhesive application. After water storage for 21 days and thermocycling (2000x, 5-55 degrees C) replicas were produced for quantitative marginal analysis in the SEM. Teeth were immersed in 0.5% basic fuchsin for 24 hours and dried. Percent dye penetration over the total marginal length was analyzed in three layers using a sequential grinding technique. Statistical analysis was performed using a two-way ANOVA. Post-hoc analyses were carried out with univariate Mann-Whitney-U-tests adjusting for multiple comparisons by a sequentially rejective test procedure (Bonferroni-Holm) at p <0.05. Both F2000 and Vitremer sandwich restorations showed better marginal adaptation than Z100 total-bond restorations with all pretreatments. Acid etching of the dentin significantly influenced the marginal adaptation of Z100 total-bond restorations and Vitremer sandwich restorations. All types of restorations showed considerable microleakage. On contaminated dentin, sandwich restorations showed better marginal integrity than total-bond restorations. Marginal adaptation did not correspond with microleakage in all groups. In conclusion, F2000/Z100 and Vitremer/Z100 sandwich restorations show better marginal adaptation than Z100 total-bond restorations in large Class II cavities with cervical margins in dentin. Microleakage cannot predictably be prevented with the sandwich technique. Sandwich restorations seem to be less sensitive to contamination with saliva and blood.
|Number of pages||10|
|Publication status||Published - 1 Jan 2001|