INTRODUCTION

The detrimental impact of resin composite polymerization shrinkage on restoration interface quality and stability has been recognized since the early use of this material and has led to the recommendation of various compensating procedures such as incremental methods,^1-5 the use of ceramic inserts,⁶ the application of base/liners,^2,3,7-9 or indirect techniques.¹⁰ The aforementioned, traditional restorative solutions shared the objective of limiting the development of marginal or internal defects, which can affect restoration behavior and longevity.

The use of incremental techniques was validated by several long-term clinical reports, demonstrating also the satisfactory clinical behavior of hybrid and microhybrid composite technology^11-13; however, these methods, which are considered the “standard of care” for managing polymerization stresses in medium to large class II cavities, all have been at times considered cumbersome and stimulated the dental industry and clinicians to look for alternative and simplified operative protocols. Various technological solutions were then investigated, including improved filler technology (blend of fillers and/or increase of filler load); improved, novel matrix structure with reduced shrinkage (ie, silorane)^14,15; use of stress-decreasing compounds within the resin matrix^16,17; changes in light-initiation technology to increase curing depth; and use of sonic vibrations and energy to favor flow and adaptation of highly filled resin composite (SonicFill, patent US7014462 B1). Today, however, little information exists about the real clinical benefit of these new technologies used alone as a simplified filling method or in combination with some of the aforesaid stress control concepts.

The true challenge for simplified restorative systems based on a bulk-filling approach or the application of a thick flowable composite base underneath a single layer of restorative material is to demonstrate both satisfactory initial quality and medium- to long-term behavior (marginal, surface, and restoration bulk integrity) when compared with conventional composite technology and clinical protocols. So far, the available short-term clinical trials do not demonstrate fully convincing performance of either silorane technology or bulk-fill techniques on permanent teeth.^18-20 The short- and medium-term performance of bulk-filling technique proved satisfactory in only two studies dealing with the treatment of primary teeth.^21,22 The quantity and consistency of clinical evidence are clearly insufficient to unconditionally recommend these new systems.

To achieve optimal long-term performance, the requirements will be first to manage polymerization stress buildup following restoration placement and then to demonstrate suitable behavior of restored teeth to repeated oral strains though appropriate mechanical properties. Actually, the reaction to functional loading and hydrolytic degradation will define the material resistance to fatigue and subsequent interface and restoration breakdown.²³ The fatigue behavior of restorations and the occurrence of adhesive or cohesive failures depend on some important material properties such as flexural strength, fracture toughness (K_Ic), and elasticity modulus (E).^24-26 It was shown, for instance, that the latest property has a crucial impact on stress development within the tooth restoration system.^2,3,7,27,28 The use of an elastic base or liner with low E modulus, acting as a stress-breaker element within the restoration, has been extensively evaluated since the first works by Davidson and coworkers^27-29 and largely validated in vitro.^30,31 When using new simplified restorative systems featuring distinctive physicochemical characteristics, the potential impact of the above mentioned parameters on restoration quality and behavior is unknown and justifies additional investigations. In consideration of a rather well-established consensus suggesting medium- to long-term observation periods (about three to five years) to discriminately appraise the clinical performance of various operative protocols and material choices,^32-35 the use of in vivo, preclinical trials such as fatigue testing^36-39 appears particularly suitable today for evaluating new, simplified restorative protocols.

The aim of this in vitro study was to test the hypothesis that low-shrinkage restorative systems have the potential to improve restoration adaptation after simulated in vitro occlusal stressing, as compared with a traditional microhybrid restorative composite applied with a classical, well-documented incremental technique. The quality of the different interfaces was also evaluated to identify the restoration's most vulnerable areas.

METHODS AND MATERIALS

Specimen Preparation

Forty freshly extracted human third molars were used for this study. Samples were collected anonymously, and their use complied with all local human subject oversight according to the Swiss Human Research Act, under article 2.

The inclusion criteria were an absence of carious lesions, a complete root formation, and no visible tooth defect resulting from the extraction. The teeth were stored in a sodium azide solution (0.2%) at 4°C until the experiment onset. For each specimen, the root length was adjusted to fit in the test chamber of the mechanical loading device (Department of Cariology, Endodontics & Pedodontics, Laboratory of Electronics of the Medicine Faculty, University of Geneva). After the specimen was properly positioned, it was fixed with light-curing composite on a metallic holder (Baltec, Balzer, Liechtenstein); then, the root base was embedded with self-curing acrylic resin to complete the tooth stabilization, leaving the last 3 mm toward the cementoenamel junction free to provide proper access to restorative procedures.

A new, original technique was used to box proximal cavities, with the aim to facilitate material placement, improve restoration adaptation, and shorten finishing steps. Then, an impression of the tooth was made prior to cavity preparation with a putty silicone (Memosil 2, Heraeus Kulzer, Hanau, Germany) placed in a 2 cm diameter cylindrical plastic tube used as tray. After setting, the tube was removed and all the silicone material overlaying the tooth occlusal surface trimmed off before slicing this impression in two parts buccolingually (Figure 1). To simulate the use of a metallic matrix such as commonly employed, aluminum foil was adapted with a burnisher to each proximal surface of the silicone impression just before tooth preparation and application of the filling material, as further described.

View Figure

• Download as Powerpoint
• Download Figure

Class II cavities (MOD) were then prepared, with the proximal margins located mesially 1.0 mm coronal to and distally 1 mm apical to the cementoenamel junction (Figure 2). The dimensions of the slightly divergent preparations were 4.0 mm in width and 2.0 mm in depth at the bottom of the proximal box and 3.0 mm in width and depth for the occlusal isthmus, all walls being only slightly divergent. The cavities were prepared using preparation diamond conical burs with round tips under profuse water spray (80 μm grain size, ISO 856/018, Strauss & CO, Ra'Anana, Israel). All enamel margins were beveled with a fine diamond flame bur (40 μm grain size, ISO 862/010, Strauss & CO) under air spray. The 40 prepared teeth were randomly assigned to one of the five study groups (one active control group and four experimental groups), corresponding to the combination of restorative materials and procedures listed in Table 1 and 2.

View Figure

• Download as Powerpoint
• Download Figure

Table 1 Combination of Products (Adhesive, Liners, and Restorative Materials) Among the Control and Four Experimental Groups

View Fullscreen

Table 2 Composition and Relevant Physical Characteristics of Composites Under Evaluation (Manufacturer's Data)

View Fullscreen

Specimen Evaluation

Before the fatigue test, as well as after completion of the loading phase, the restoration's margins were cleaned with fine pumice before etching using diluted H₃PO₄ gel to remove finishing/polishing smear and improve the readability of the replicas under scanning electron microscopy (SEM; the solution was prepared with 35% H₃PO₄ gel mixed with three times its volume of distilled water). The restoration surfaces and margins were then treated by gentle brushing of this solution for four seconds, followed by thorough water rinsing. Impression of the samples was made with polyvinylsiloxane (President Light Body, Coltene/Whaledent AG, Alstätten, Switzerland), which served for the fabrication of gold-sputtered epoxy resin replicas (Epofix, Struers, Rødrove, Denmark).

The proximal mesial and distal aspects of the restorations were examined while the occlusal adaptation was not taken into consideration. The approximal mixed adaptation (Appr. M) refers to the entire mesial restoration adaptation (including both enamel and dentin margins). The cervical dentin adaptation (Cerv. D) refers to the specific cervical dentin adaptation on the mesial preparation side. The approximal enamel adaptation (Appr. E) refers to the distal, full enamel restoration margin adaptation. Each tooth restoration interface was analyzed quantitatively by SEM (Digital SEM XL20, Philips, Eindhoven, the Netherlands) using a standardized evaluation method.^41,42 The restoration margins were observed at a standard 200× magnification, and each margin segment, or subsegment, was delimitated by a digital marker and attributed a quality criteria, as described further (Figure 3). A dedicated software then computed the respective percentages of continuous or defective margin for the three different areas under evaluation applying those quality criteria: continuity or overfilling, underfilling, marginal opening, marginal restoration, and tooth fracture (Figure 1). When necessary for the assessment accuracy, higher magnifications were used (up to 1000×). The results for the restoration marginal adaptation before and following the loading test were expressed as percentages of continuity; the other parameters that describe defective margin sections were pooled together and accounted for noncontinuous adaptation and were not analyzed individually. Percentages were calculated as the ratio between the cumulated distance of all segments with a continuous margin and the whole interface length. The restoration occlusal adaptation was not assessed.

View Figure

• Download as Powerpoint
• Download Figure

All results of the SEM analysis were submitted to a parametric statistical analysis. An analysis of variance and Bonferroni post hoc test served for comparing the intergroup marginal continuity percentages for the different restoration areas, before and after the loading test (Instat, GraphPad Software, La Jolla, CA, USA). A paired t test served to compare the intragroup marginal adaptation percentages between pre- and postloading conditions (Instat, GraphPad Software). All tests were carried out at a 5% level of significance.

RESULTS

The study results, expressed as percentages of continuous marginal adaptation before and after loading, are presented in Table 3 together with their statistical analysis.

Table 3 Percentages of Continuous Marginal Adaptation (±SD) for the Four Segments Under Evaluation, Pre- and Postcycling Loading^a

View Fullscreen

DISCUSSION

Phenomena such as nanoleakage, leakage, pulpal complications, and secondary caries, which are induced by interface breakdown, account for a significant part of clinical failures observed in all types of direct posterior restorations.^11-13 Then, evaluating the behavior of adhesive restorations with natural tissues under simulated function, pulpal pressure, and moist environment helps in approaching the reaction of a restoration to the most important oral cavity strains and to monitor the tooth-composite interface stability and degradation as well.^36-39 This in vitro, preclinical research approach is well established and has been used previously in numerous studies evaluating the in vitro quality of class II restorations.^30,31,39

Overall, the materials and restorative techniques under evaluation presented satisfactory adaptation to approximal enamel (distal) and to mixed enamel-dentin margins (mesial) before and after loading (>90% continuous margins), with the exception of SonicFill and Tetric. Actually, the restorations made with SonicFill showed more postloading marginal microfractures in plain enamel margins whereas the restorations made with Tetric showed a high proportion of dentin marginal gaps, pre- and postloading. The higher occurrence of marginal enamel microfractures reported for SonicFill might be related to the higher stiffness of this product (highest E modulus among tested products; see Table 1), which likely increases stress transmission to the restoration margins, especially when used with a bulk-filling approach. The possible clinical impact of such adverse finding, commonly observed in previous in vitro fatigue studies, has not been clarified.^9,43

The most critical margin area was as usual the cervical dentin, where the multilayered restorations made of a traditional microhybrid (Tetric) and a so-called “low-shrinkage” microhybrid (ELS) used alone or in combination with a flowable resin composite liner (ELS + ELSflow) did not perform satisfactorily, as shown by significantly lower continuity proportions in pre- (Tetric only) and postloading conditions. In studies evaluating marginal adaptation in similar in vitro conditions, Tetric⁴⁴ and Tetric Ceram,⁴⁵ which are both microhybrids, and Tetric EvoCeram (nanohybrid)⁴⁴ placed with other adhesives presented percentages of postloading continuity varying from 83.3% to 94.6% in enamel and from 56.2% to 74.6% in dentin, which is markedly higher than in the present study for dentin adaptation (18.66%). Kwon and others⁴⁶ using the same adhesive with Tetric Ceram obtained 78.7% of continuity in mixed dentin and enamel margins, which is nearly identical to the result of the present study, with 77.0% continuous margins for the multilayered Tetric class II restorations. When considering the adaptation to dentin, the performance of Tetric then fell below the range of published data for multilayered direct composite restorations tested in similar laboratory environments. As the present study protocol involved a higher cycling load (100N instead of 49-50N or varying forces from 50 to 100N), or a higher number of cycles (one million instead of 100,000 to 600,000), it could have had a more damaging and discriminative effect on dentin adaptation, although some unidentified confounding factor could also account for such surprisingly low performance.

In two recent studies, SDR flow combined with Ceram-X in class II restorations submitted to a similar fatigue test (but with a reduced number of cycles) presented postloading percentages of continuity in dentin of 50.3% and 64.9%, depending on the adhesive tested,⁴³ and 42.9%,⁴⁷ with those values being rather close to those reported in the present study. SonicFill presented percentages of continuity in dentin, postloading, of 51.1% in the present study and 61.7% in the other study.⁴⁷ While these percentages can be considered apparently satisfactory, such a proportion of continuous margin remains largely inferior to the best combination of adhesive, flowable resin liner, and restorative composite, peaking at 90.8% in a similar test environment for indirect composite restorations⁹ or to classical multilayered full-composite restorations with continuous adaptation in dentin reaching 74.6% to 78%.⁴⁸

The E modulus and volumetric shrinkage are of particular interest^49,50 as both properties are known to significantly affect polymerization stresses. Actually, for a given volumetric shrinkage, a lower elasticity modulus will help to reduce stresses. On the other hand, the material's stiffness (as defined by its E modulus) also influences restoration adaptation following loading because of distinct deformation and stress absorption capacity.^9,44,48 In this study, the SDR group (SDR flow E modulus=5.5 GPa) presented a better dentin and overall adaptation than the group with the ELSflow liner (E modulus=8 Gpa); considering a volumetric shrinkage within the same range, the material's stiffness is potentially the influential property. Considering again the interaction between shrinkage stress and E modulus, the similar results obtained with the ELS restorative material alone or combined with the ELSflow become more logical and suggest that using a lining with an almost similar E modulus (ELSflow and ELS E modulus are, respectively, 8 and 9 GPa) is unlikely to give any biomechanical advantage to the restoration's adaptation, although it facilitates and quickens clinical procedures.

Another influential parameter is the respective cavity and increment/layer C-factor (configuration factor),³⁶ which justified many evidenced-based restorative protocols.^30,31 The rationale for placing a flowable liner underneath direct class II composite restorations is then based on a low C-factor of this first layer, promoting reduced stress buildup at the cavity-restoration interface, combined also with the superior stress absorption capability of a flowable resin composite.^46,51-54 In consideration of the C-factor, simplified layering techniques (bulk-filling or extended flow base) do, however, not show any particular advantage.

In an attempt to summarize the main parameters that explain the present study's findings, the better performance of the SonicFill and SDR/Ceram-X groups in cervical dentin is potentially linked to, respectively, the low volumetric shrinkage of SonicFill and the low E modulus of SDR flow base/Ceram-X systems. As well, the rather unsatisfactory results of the TET and ELS and combined ELS/ELSflow groups are to be attributed to the higher volumetric shrinkage and E modulus of Tetric and the intermediate E modulus and medium/high volumetric shrinkage of ELS/ELSflow materials, which overcame the theoretical advantage of using a multilayering approach.

The present study tested materials and filling methods in medium-size cavities, which could have been more favorable to a low E modulus material (ie, SDR flow) because of the less critical reinforcement role played by the restoration. In larger cavities or in biomechanically compromised teeth, such nonvital teeth, long-term functional loading, and higher occlusal forces (ie, bruxism) could have a more detrimental effect on restoration adaptation because of increased tooth structure deformation. More research is then needed to confirm the relationship between cavity size, material stiffness, and long-term restoration biomechanical behavior.

CONCLUSIONS

An in vitro evaluation of marginal adaptation following medium-term functional loading simulation (one million cycles) was used for evaluating new filling methods and materials proposed for class II restorations.

In this preclinical test, medium-size class II restorations made with a traditional layering approach and flowable composite resin liner or simplified filling methods presented satisfactory adaptation to proximal enamel, whereas in cervical dentin, the bulk-filling technique (SonicFill) and extended flow base (SDR flow + Ceram-X) showed the best adaptation. Reported values of continuity were, however, inferior to many direct or indirect composite restorations using traditional microhybrid composites, as reported in previous, similar in vitro trials. As some of the new, simplified restorative systems exhibit singular composition and physical characteristics, extended fatigue tests and clinical trials will be needed to assess their performance, in particular in larger cavities and with higher functional stresses. The use of new materials for bulk filling or a simplified restorative protocol cannot yet be unconditionally recommended, based on their in vitro performance.