Abstract
Objectives
Polymerization of composite restorations causes shrinkage, which deforms and thus stresses restored teeth. This shrinkage deformation, however, has been shown to decrease over time. The objective was to investigate whether this reduction was caused by hygroscopic expansion or stress relaxation of the composite/tooth complex.
Methods
Extracted molars were mounted in rigid stainless steel rings with four spherical reference areas. Twelve molars were prepared with large mesio-occluso-distal slots, etched, bonded, and restored with a composite material (Filtek Supreme, 3M ESPE) in two horizontal layers. Ten intact molars were the controls. The teeth were stored either in deionized water or silicone oil. They were scanned after preparation (baseline), restoration (0-week), and after 1, 2, and 4 weeks storage. Scanned tooth surfaces were aligned with the baseline using the unchanged reference areas. Cuspal flexure was calculated from lingual and buccal surface deformation. To verify that the restorations had remained bonded, dye penetration at the interfaces was assessed using basic fuchsin dye. Statistical assessment was done by ANOVA followed by Student–Newman–Keuls post hoc test ( p = 0.05).
Results
Substantial cuspal contraction was found for restored teeth after the composite was cured (13–14 μm cuspal flexure). After 4 weeks cuspal contraction decreased significantly for restored teeth stored in water (7.3 ± 3.2) but not for those stored in silicone oil (11.4 ± 5.0). Dye penetration of the occlusal interface was minimal in both groups (106 ± 87 and 21 ± 28 μm in water and silicone oil, respectively).
Significance
The results suggest that hygroscopic expansion was the main mechanism for shrinkage stress compensation.
1
Introduction
Polymerization of resin-based composites results in densification and thus volumetric contraction . If a polymerizing composite restoration is bonded to the walls of a cavity, such material shrinkage is hindered, which may result in residual stresses to develop in the composite, across the adhesive interface, and in the restored tooth.
Shrinkage stress has been a concern since introduction of resin-based composites to dentistry because it is thought to be associated with a variety of clinical issues, including post-operative sensitivity, adhesive bond failure, and enamel crack propagation . Although shrinkage stresses themselves cannot be directly measured, they are often evident in cuspal flexure . However, tooth deformation has been shown to decrease over time, suggesting an accompanying reduction of shrinkage stresses . The reduction in cuspal flexure can be caused by various mechanisms, including debonding of the adhesive interface , stress relaxation of the resin-based materials and/or tooth tissues , and/or shrinkage compensation by hygroscopic expansion .
Most polymerization shrinkage research concerns the initial stress development during the restorative procedure when a composite shrinks and its elastic modulus develops. Although this initial shrinkage stress may be important for choosing restorative materials and techniques, how these residual stresses evolve further will ultimately determine the clinical impact of shrinkage and if shrinkage will affect the longevity of a restored tooth. To better understand the long-term effect of polymerization shrinkage stresses in teeth restored with an adhesive resin composite, the objective of this study was to examine the contribution of hygroscopic expansion and stress relaxation as stress relief mechanisms. Since stress relaxation and hygroscopic expansion should be expected to take place simultaneously, the contribution of stress relaxation was isolated by using silicone oil as a non-aqueous non-desiccating environment . Dye penetration tests were used to verify that the restorations had remained bonded during the experiments.
2
Materials and methods
Eleven pairs of extracted molars (Institutional Review Board exempt category study number 11-01209-XM) were mounted tightly in rigid stainless steel rings ( Fig. 1 ). Matched molar pairs were selected based on arch, type and shape such as number of cusps and size. The average difference in bucco-lingual width between pairs was less than 2.5%. Enamel was etched with 37% phosphoric acid solution so that it could be scanned by the optical scanner (Comet xS, Steinbichler Optotechnik GmbH, Neubeuern, Germany) ( Fig. 1 ).
Mesio-occluso-distal slots, 4 mm wide and 4 mm deep were prepared in six pairs of molars using a high-speed handpiece with #245 carbide bur with copious water coolant. The average (±standard deviation) remaining wall thickness of the prepared teeth was 2.7 ± 0.4 mm. Ten minutes after preparation, the prepared molars were scanned from eight different angles with a 60-μm lateral resolution and 5-μm accuracy. Between the different experimental steps, teeth were kept hydrated to prevent desiccation. The scan of the tooth with preparation was used as the baseline. After the preparation had been scanned, it was etched with phosphoric acid gel (Lot 9NK; 3M ESPE, St Paul, MN, USA) for 15 s, rinsed 15 s, and blot dried. Two layers of Adper SingleBond Plus adhesive (3M ESPE) were applied and light cured for 10 s with a quartz–tungsten–halogen curing unit (XL 3000, 3M ESPE) that had an output of 450 mW/cm 2 measured with a curing radiometer (Model 100, Demetron Research Corp., Danbury, CT, USA). The cavities were restored with Filtek Supreme Plus Universal Restorative composite (3M ESPE) in two horizontal layers of 2 mm thickness. See Table 1 for the adhesive and restorative materials description. Due to the size of the preparation, the mesial and distal halves of each layer were light cured for 40 s from the occlusal direction, thus a total of 80 s per layer. The restoration surfaces were wiped with alcohol pads to remove the air-inhibited layer but were not polished. The restored molars were scanned 10 min after restoration (0-week). One molar from each pair was randomly selected to be stored in silicone oil (polydimethylsiloxane, Acros Organics, Geel, Belgium) as a non-aqueous medium while the other was stored in deionized water. There was no statistically significant difference in remaining wall thickness between the restored molar groups stored in silicone oil or water ( t -test, p = 0.528). Five pairs of intact molars were submitted to the same scanning and storage protocols and served as controls. All groups were stored at room temperature and scanned again after 1, 2, and 4 weeks.
Type | Product a | Description and composition b | Shade | Lot |
---|---|---|---|---|
Resin composite | Filtek Supreme Plus Universal Restorative | Visible-light activated nanocomposite with a BIS-GMA, BIS-EMA, UDMA, and TEGDMA resin system, containing non-agglomerated/non-aggregated 20 nm nanosilica fillers and loosely bound agglomerates of 5–20 nm zirconia/silica nanoclusters with particle sizes ranging between 0.6 and 1.4 μm; filler loading 78.5% by weight | A2D | 8BE |
Adhesive | Adper Single Bond Plus Adhesive | Total etch, visible-light activated adhesive composed of BIS-GMA, HEMA, dimethacrylates, ethanol, water, methacrylate functional copolymer of polyacrylic and polyitaconic acids, a photoinitiator system, and 5 nm silane-treated spherical silica particles; filler loading 10% by weight | – | 365812 |
After each scan, the digitized teeth were accurately aligned with their baselines, using the four stainless steel spheres embedded in the mounting rings as (unchanging) reference areas ( Fig. 2 A). The fitting was achieved by minimizing the root-mean-square differences between the spherical stainless steel surfaces using Cumulus software (© Regents of the University of Minnesota). The aligned tooth models were exported to a custom software program (CuspFlex) where buccal and lingual surface areas were selected for which tooth deformation was determined ( Fig. 2 B and C). Buccal and lingual cuspal deflections were calculated from the volume change/area. The results were statistically analyzed using ANOVA followed by Student–Newman–Keuls post hoc test ( p = 0.05).
Dye penetration of the occlusal interface of the restorations was assessed after the cusp flexure experiments had been finished to verify the quality of the bonding. Restoration margins were polished using composite finishing burs (LogicSet, Axis Dental Corp, Coppell, TX). Root apices were obstructed with utility wax and the roots were painted with nail polish. The teeth were immersed in 0.5 wt% basic fuchsin solution for 20 h, embedded in acrylic resin, and sectioned bucco-lingually every 1 mm. Each restored molar yielded 6–8 sections. The images of the cross-sections were recorded using a stereomicroscope and CCD camera (SZX16 & UC30, Olympus, Tokyo, Japan). Dye penetration distances were measured along the buccal and lingual interfaces at the occlusal margins of the restoration by two evaluators using image analysis software (Stream Basic, Olympus Soft Imaging Solution GmbH, Münster, Germany). If the measurements between the evaluators differed more than 10%, a consensus value was reached. The dye penetration distance for each tooth was the average from the buccal and lingual values of all its sections. Mean dye penetration distance was calculated from all six restored teeth per group. The t -test was used to test if there was any statistically significant difference in dye penetration between the two storage media.
2
Materials and methods
Eleven pairs of extracted molars (Institutional Review Board exempt category study number 11-01209-XM) were mounted tightly in rigid stainless steel rings ( Fig. 1 ). Matched molar pairs were selected based on arch, type and shape such as number of cusps and size. The average difference in bucco-lingual width between pairs was less than 2.5%. Enamel was etched with 37% phosphoric acid solution so that it could be scanned by the optical scanner (Comet xS, Steinbichler Optotechnik GmbH, Neubeuern, Germany) ( Fig. 1 ).
Mesio-occluso-distal slots, 4 mm wide and 4 mm deep were prepared in six pairs of molars using a high-speed handpiece with #245 carbide bur with copious water coolant. The average (±standard deviation) remaining wall thickness of the prepared teeth was 2.7 ± 0.4 mm. Ten minutes after preparation, the prepared molars were scanned from eight different angles with a 60-μm lateral resolution and 5-μm accuracy. Between the different experimental steps, teeth were kept hydrated to prevent desiccation. The scan of the tooth with preparation was used as the baseline. After the preparation had been scanned, it was etched with phosphoric acid gel (Lot 9NK; 3M ESPE, St Paul, MN, USA) for 15 s, rinsed 15 s, and blot dried. Two layers of Adper SingleBond Plus adhesive (3M ESPE) were applied and light cured for 10 s with a quartz–tungsten–halogen curing unit (XL 3000, 3M ESPE) that had an output of 450 mW/cm 2 measured with a curing radiometer (Model 100, Demetron Research Corp., Danbury, CT, USA). The cavities were restored with Filtek Supreme Plus Universal Restorative composite (3M ESPE) in two horizontal layers of 2 mm thickness. See Table 1 for the adhesive and restorative materials description. Due to the size of the preparation, the mesial and distal halves of each layer were light cured for 40 s from the occlusal direction, thus a total of 80 s per layer. The restoration surfaces were wiped with alcohol pads to remove the air-inhibited layer but were not polished. The restored molars were scanned 10 min after restoration (0-week). One molar from each pair was randomly selected to be stored in silicone oil (polydimethylsiloxane, Acros Organics, Geel, Belgium) as a non-aqueous medium while the other was stored in deionized water. There was no statistically significant difference in remaining wall thickness between the restored molar groups stored in silicone oil or water ( t -test, p = 0.528). Five pairs of intact molars were submitted to the same scanning and storage protocols and served as controls. All groups were stored at room temperature and scanned again after 1, 2, and 4 weeks.