An In Vitro Evaluation of Leakage of Two Etch and Rinse and Two Self-Etch Adhesives after Thermocycling

Our experiment evaluated the microleakage in resin composite restorations bonded to dental tissues with different adhesive systems. 40 class V cavities were prepared on the facial and lingual surfaces of each tooth with coronal margins in enamel and apical margins in cementum (root dentin). The teeth were restored with Z100 resin composite bonded with different adhesive systems: Scotchbond Multipurpose (SBMP), a 3-step Etch and Rinse adhesive, Adper Scotchbond 1 XT (SB1), a 2-step Etch and Rinse adhesive, AdheSE One (ADSE-1), a 1-step Self-Etch adhesive, and AdheSE (ADSE), a 2-step Self-Etch adhesive. Teeth were thermocycled and immersed in 50% silver nitrate solution. When both interfaces were considered, SBMP has exhibited significantly less microleakage than other adhesive systems (resp., for SB1, ADSE-1 and ADSE, P = 0.0007, P < 0.0001 and P < 0.0001). When enamel and dentin interfaces were evaluated separately, (1) for the Self-Etch adhesives, microleakage was found greater at enamel than at dentin interfaces (for ADSE, P = 0.024 and for ADSE-1, P < 0.0001); (2) for the Etch and Rinse adhesive systems, there was no significant difference between enamel and dentin interfaces; (3) SBMP was found significantly better than other adhesives both at enamel and dentin interfaces. In our experiment Etch and Rinse adhesives remain better than Self-Etch adhesives at enamel interface. In addition, there was no statistical difference between 1-step (ADSE-1) and 2-step (ADSE) Self-Etch adhesives.

All the tested adhesives were used according to the manufacturer's instructions. Immediately after bonding procedures, the cavities were filled with two oblique increments of a microhybrid composite (Z100, 3M ESPE AG, Dental products, Seefeld, Germany). The photopolymerization was carried out for all materials with a halogen lamp (XL 3000, 3M ESPE AG, Dental products, Seefeld, Germany). Composite restorations were polished by means of diamond drills and disks (Hawe Neos Dental, Bioggio, Switzerland). The polishing was carried out under a spray of water. After that, all the specimens were immersed in a saline solution for twelve weeks (in a refrigerator at 5 • C). Thereafter, the teeth were thermocycled for 800 cycles (5 • C-55 • C) for 22 hours. After thermocycling, the teeth were immersed in a 50% silver nitrate solution (for 6 hours) and in a 25% vitamin C solution for 10 minutes (pH about 2) [25,26]. After immersion, three grooves (3 mm depth, 1 mm width) were drilled with a diamond bur in each restoration to obtain four surfaces of observation. The interfaces that occurred between the teeth and the filling was described in our previous studies [25,26]. Briefly, the cylindrical diamond drill (0.9 mm diameter) was placed perpendicular to the composite restoration. Three grooves 3 mm deep and 1 mm wide were cut on each restoration: one at the mesial margin, one at the distal margin, and one right in the middle of the filling (Figure 1). These preparations yielded four evaluating surfaces for each composite restoration (Figure 2), for a total of 160 viewing surfaces for all tested adhesives. Each surface allowed one observation in enamel and one in dentin (lecture areas), for a total of 320 observations (160 in enamel and 160 in dentin).
Each section was examined by twofold magnification by means of an optic microscope (Carl Zeiss, SAS, Oberkochen, Germany). Each tooth was observed twice by the same operator (blinded test).
Arbitrarily, the evaluation of leakage was made with a 6point severity scale (Figure 3, Table 2) [25].
We have postulated that higher scores of microleakage (scores 3, 4, and 5) after thermocycling would be responsible for clinical failure of the bonding (Figure 4).

Surfaces for observations
Composite restoration before grooves

Grooves
Zone of leakage score evaluation

Statistical Analysis.
Results are expressed as means ± standard deviations (± SDs). Microleakage scores were analyzed by means of generalized linear mixed models (GLMMs) assuming an ordinal logistic link function. Covariates in the model were (1) adhesive systems and (2) interface (enamel or dentin). The model also accounts for repeated measurements on the various teeth. All the results were considered to be significant at the 5% critical level (P < 0.05). Statistical calculations were made using the SAS 9.1 (version 8.2 for Windows) package.

Results
Microleakage mean score calculation for each tested adhesive system was analyzed by statistical model, which takes repeated evaluations into account (4 observations for each interface, enamel, or dentin). Therefore, all the observed scores of microleakage for each adhesive at enamel or at dentin interface are not displayed. The mean scores of microleakage for all tested adhesive systems are shown in Table 3.
In our study, SBMP was significantly different from other adhesives: SBMP has shown a lower mean score of microleakage (0.30 ± 0.49) than other tested adhesives (P = 0.0007 for SB-1 and P < 0.0001 for the other tested adhesives). Table 4 reports the statistical comparison between the mean scores of microleakage of the 4 tested adhesives.
As seen in Table 4, there was no statistical difference between SB1 and ADSE (P = 0.0799), neither between SB1 and ADSE-1 (P = 0.072) nor between ADSE and ADSE-1 (P = 0.96). Table 5 shows the mean scores of microleakage for the 4 tested adhesives at enamel and dentin interfaces.
For ADSE and ADSE-1, the mean scores of microleakage were significantly lower at dentin than at enamel interfaces.

Discussion
For the past few years, composite has become current restorative material and today it often replaces amalgam restorations in posterior teeth [28][29][30][31]. Nevertheless, restorative composite is not able to bond to dental tissues. Therefore, the use of an adhesive system is always required. As result of numerous advances in adhesive technology and adhesion knowledge, there are many adhesive systems available on the market. To avoid confusing and incorrect uses of the adhesives, Professor Bart Van Meerbeek has proposed a classification according to different adhesion strategies and adhesives: the Etch and Rinse (ER) adhesive systems, the Self-Etch (SE) adhesive systems, and the glass ionomer adhesive systems [6,7,32]. The ER adhesives always involve the use of phosphoric acid, which permits demineralization of the dental tissues and, after rinsing, a complete elimination of the smear layer. Therefore, in the course of the ER adhesion   strategy, the adhesive resin (bonding) is applied in a different clinical step: the demineralization and the hybridization of dental substrate appear consecutively. On the contrary, with the SE adhesives the demineralization and the impregnation of the adhesive into the enamel-dentin support appear simultaneously. The demineralization process results from the acidic monomers, which are components of the adhesive system. Therefore, the SE adhesive must not be rinsed. There are currently 4 different types of SE adhesives, which are indexed according to their pH value: the ultramild SE (pH about 2.5), the mild SE (pH about 2), the intermediary International Journal of Dentistry 5 strong SE (pH about 1.5), and the strong SE (pH < 1) [6][7][8]32]. On the enamel, for both ER and SE adhesive systems, bonding to the tissue is essentially micromechanical. On the dentin, for the ER adhesives, the mechanisms of adhesion are mainly micro-mechanical because the phosphoric acid is a very strong acid (pH about 0.5). Phosphoric acid completely dissolves the mineral and so, the collagen fibers are totally exposed after etching. For the SE adhesives, the adhesion to the dentin is both micro-mechanical and chemical [6][7][8]: the self-etch monomers are often less acidic than phosphoric acid and then some minerals remain attached to the collagen fibers, permitting chemical links between dental substrate and functional groups of the adhesive monomers.
Nevertheless, for some authors, 2-step mild and ultramild SE adhesives can give comparable results than those obtained by some 2-step ER adhesives and also, by some 3-step ER adhesive systems [6-10, 32, 34, 36, 37]. In fact, our results have shown that ADSE (2-step SE) and SB1 (2-step ER) have statistically comparable mean scores of microleakage.
Concerning the 1-step SE adhesives, some in vitro studies have shown their poor performances [7,34,46]. Our results do not confirm this fact: there is no statistical difference between the mean scores of microleakage of ADSE (a 2step SE) and its simplified clinical version, ADSE-1 (a 1-step SE). In addition, the mean scores of microleakage of these two mild SE adhesives (pH about 2) are lower at the dentin interface than at the enamel interface. These observations agree with data from the literature: ADSE and ADSE-1 are not efficient to create a sufficient micro-mechanical retention at the enamel surface [6,7,16,34,43]. Nevertheless, at the dentin surface, these mild SE adhesives are able to create a partial demineralization of this tissue to allow a micromechanical adhesion [6,7]. In addition, some functional monomers of these SE adhesives might form chemical bonds with the calcium of the residual hydroxyapatite crystals linked to the collagen fibers [7,32,[38][39][40]. The chemical bonds between ADSE functional monomers have not been clearly identified yet, but this adhesive has given good results in our study, like in the study of Bradna et al. [10].

Conclusion
In this study, confirming previous studies about marginal microleakage of the ER adhesive systems, SBMP, a 3-step ER adhesive, has significantly less microleakage comparing to other adhesive systems and can be considered like a reference adhesive. The parameters of this experiment (hydrolysis and thermocycling) have shown the good in vitro behaviour of SBMP. Therefore, we can expect that this ER adhesive will be clinically satisfying. In fact, this adhesive has been widely used for many years and their performances have seemed good. The 2-step ER adhesive that was tested in our study has shown a significantly greater mean score of microleakage than the tested 3-step ER adhesive system, but all the tested adhesives showed minimal leakage.
In the limits of our study, ADSE and ADSE-1 show poor microleakage, particularly on the dentin. Nevertheless, we suggest these mild SE adhesives can be used when the margins of the cavity are located on dentin and/or using phosphoric acid only on the enamel margin in order to optimize micro-mechanical interlocking.