Correlation between polymerization stress and interfacial integrity of composites restorations assessed by different in vitro tests

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.intl.elsevierhealth.com/journals/dema

Correlation between polymerization stress and interfacial integrity of composites restorations assessed by different in vitro tests Letícia Cristina Cidreira Boaro a,b,∗ , Nívea Regina Fróes-Salgado b , Vinicius Edwardo Souza Gajewski b , Aline A. Bicalho c , Andrea Dolores Correa M. Valdivia c , Carlos José Soares c , Walter Gomes Miranda Júnior b , Roberto Ruggiero Braga b a b c

Universidade de Santo Amaro, São Paulo, SP, Brazil Universidade de São Paulo, São Paulo, SP, Brazil Universidade Federal de Uberlândia, Uberlândia, MG, Brazil

a r t i c l e

i n f o

a b s t r a c t

Article history:

Objective. to correlate polymerization stress data obtained under two compliance conditions

Received 16 February 2013

with those from different interfacial quality tests.

Received in revised form

Methods. Six commercial composites were tested (Filtek Z250/3M ESPE, Heliomolar/Ivoclar

10 December 2013

Vivadent, Aelite LS Posterior/Bisco, Filtek Supreme/3M ESPE, ELS/Saremco and Venus Dia-

Accepted 21 May 2014

mond/Heraeus Kulzer). Bond strength (BS) was evaluated by push-out test on slices of bovine dentin (2-mm thick) with tapered cavities. For microleakage (ML) and gap analysis, cylindrical cavities in bovine incisors (4-mm diameter and 1.5-mm height) were restored and epoxy

Keywords:

replicas of the cavo-surface margins were prepared for analysis under scanning electron

Polymerization stress

microscopy (200×). The same specimens were submitted to a microleakage protocol using

Composites

AgNO3 as tracer. After sectioned twice perpendicularly, ML was determined under a stereo-

Bond strength

microscope (60×). Polymerization stress (PS, n = 5) was determined by the insertion of the

Marginal gap

composite (h = 1.5 mm) between poly(methyl methacrylate), PMMA, or glass rods (Ø = 4 mm) attached to a universal testing machine. Data were analyzed using Kruskal–Wallis (ML and gaps), and ANOVA/Tukey (BS and PS, ˛ = 5%). Pearson’s correlation test was used to verify correlations between stress and interfacial quality. Results. BS varied from 4.7 to 7.9 MPa. Average ML data ranged from 0.34 to 0.89 mm. Maximum ML varied from 0.61 to 1.34 mm. Gap incidence varied from 13 to 47%. PS ranged from 2.5 to 4.4 MPa in PMMA, and between 2.1 and 8.2 in glass. Statistically significant correlations were observed between stress and interfacial quality, except between BS and PS in glass. These correlations were stronger when PMMA was used as bonding substrate.

∗ Corresponding author at: Departamento de Biomateriais e Biologia Oral da FOUSP, Av. Prof. Lineu Prestes, 2227, São Paulo, SP 05508-000, Brazil. Tel.: +55 11 3091 7840x224; fax: +55 11 3091 7840x201. E-mail addresses: [email protected], [email protected] (L.C.C. Boaro). http://dx.doi.org/10.1016/j.dental.2014.05.011 0109-5641/© 2014 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

985

Conclusions. PS data obtained using a high compliance testing system showed a stronger correlation with “in vitro” interfacial integrity results, compared to data from a low compliance system. © 2014 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

According to clinical studies, drawbacks such as postoperative sensitivity, marginal discoloration and possibly secondary caries are often associated with loss of marginal integrity in composite restorations [1–3]. One of the possible causes for interfacial debonding is polymerization stress. When composites polymerize confined in a cavity preparation, shrinkage associated with the development of modulus of elasticity generate stresses in the tooth/restoration interface, which may lead to debonding [4]. Several research groups have focused on developing mechanical tests to quantify polymerization stress [5–8]. In the most commonly used test, the composite is inserted and polymerized between two flat surfaces of glass, metal or poly(methyl methacrylate), PMMA, rods attached to an universal testing machine [9–16]. The load cell records the contraction force exerted by the composite on the substrate during polymerization and the nominal stress is calculated, in MPa, by dividing this value by the cross sectional area of the rod. This method has been widely used to compare commercial [17] and experimental composites [14,18], photoactivation methods [19] and to evaluate several factors associated with stress development [16]. Some studies have correlated the stress values from mechanical tests with interfacial integrity, noting that microleakage and cuspal deflection increase proportionally with increasing stress [4,20,21]. However, no relationship was found between stress and gap formation in porcelain inlays [22]. A study evaluating polymerization stress as a function of photoactivation methods observed that modulated photoactivation results in lower stress values, leading to higher bond strengths [19]. The studies mentioned above have in common the fact that stress was determined in low compliance systems, using glass as bonding substrate for the composite. However, the system’s compliance has great influence on ranking materials regarding stress magnitude [23,24]. The lower the compliance of the testing system, the lower is its ability to elongate and relief the stress. Consequently, the recorded value is higher. In the past few years, bonding substrates with lower modulus of elasticity have been used polymerization stress testing [11,12,23,24]. Even assuming that data from mechanical tests cannot be extrapolated to the clinic, a question arises regarding which system would be more closely related to the interfacial quality of composite restorations. It is possible that the use of low compliance testing systems could overestimate the stress values, in comparison with those found in high compliance conditions, more akin to the behavior of a prepared tooth. Estimating the compliance of the tooth in a clinical situation is nearly impossible. The stiffness of the dental tissues varies

among teeth and even in the same tooth there is a substantial difference in stiffness between enamel and dentin. But even being that complex, the tooth cannot be considered as a rigid system. In fact, several studies have shown that polymerization shrinkage could lead to tooth deformation [25,26]. In a previous study, several commercial composites ranked similarly for microleakage and stress values obtained in a high compliance system, but the same did not apply to stress data obtained in a low compliance system [23]. Considering the above, it is important to verify if data from polymerization stress tests can be correlated with results from interfacial quality tests, namely, bond strength, microleakage and gap formation. The null hypothesis was that the polymerization stress values shows no correlation to interfacial integrity, regardless of the system compliance. Additionally, a second null hypothesis was tested, stating the compliance of the testing system did not influence polymerization stress values.

2.

Materials and methods

Six dimethacrylate-based commercial composites shade A3 were tested (Table 1). Three of them (Heliomolar, Filtek Supreme and Filtek Z250) were chosen based on their filler content (by volume). The other three (Venus Diamond, ELS and Aelite LS) are considered as “low shrinkage” or “low stress” materials by the respective manufacturers. Venus Diamond has TCD-urethane in its composition, in addition to conventional dimethacrylates, while ELS has no diluent monomer (TEGDMA) and Aelite LS has a very high filler content. Elastic moduli (determined by three point bending test) and postgel shrinkage (determined by the strain-gage method) were obtained in a previous study, and correspond to the values recorded 10 min after phtoactivation using the same irradiance and radiant exposure adopted in the present study [27].

2.1.

Push-out bond strength

Bovine incisors (n = 15) had their crowns removed at the cement-enamel junction with a diamond disc under refrigeration. The buccal surface was flattened with wet sandpaper until the enamel was completely removed. The lingual surface was sectioned using a diamond disc (Isomet 1000, Buehler Ltd., Lake Bluff, IL, USA) to obtain a slice with 2 mm thickness. Tapered cavities with 2.9-mm diameter on the buccal surface and 3.5 mm diameter on the lingual surface were prepared using cylindrical and truncated cone diamond burs. The cavity walls were etched with 37% phosphoric acid for 15 s and then rinsed in running water for 15 s. Excess water was removed with short air blasts, leaving the surface visibly moist. Two layers of an one-bottle adhesive system

986

0.43 (0.02) 0.35 (0.02) 0.64 (0.07) 0.52 (0.04) 0.39 (0.03) 0.51 (0.04)

3.1 (0.3) 2.0 (0.2) 6.0 (0.7) 5.6 (0.6) 4.5 (0.3) 9.3 (0.7)

(Single Bond 2, 3M ESPE) were applied and photoactivated with a radiant exposure of 12 J/cm2 (400 mW/cm2 × 30 s – VIP Jr, Bisco, Schaumburg, IL, USA). The tooth was placed on a mylar tape over a glass slab, with the buccal surface facing up. After inserting the composite in bulk, a second mylar strip was placed on the buccal surface of the restoration and the curing tip was placed in contact therewith, so the light was directed from buccal to lingual surfaces. The composite was light cured with a radiant exposure of 18 J/cm2 (570 mW/cm2 × 32 s). Specimens were stored for 24 h in distilled water at 37 ◦ C. Both buccal and lingual surfaces were slightly ground with finishing discs (Soft-Lex, 3M ESPE). For the push-out test, the specimen was placed on a stainless steel base under the actuator of a universal testing machine (Instron 5565, Canton, MA EUA). The smallest radius (buccal surface) was placed in contact with a 2.5-mm diameter stainless steel tip, connected to the load cell. This tip applied a compressive force (cross-head speed: 0.5 mm/min) on the composite surface until the rupture of the bonded interface. Values in MPa were obtained by dividing the maximum force (N) by the bonded area of the specimen (in mm2 ). The bonded area was calculated by the formula of the lateral area of the truncated cone:

BisGMA, UDMA, D3 MA BisGMA, BisEMA BisGMA, BisEMA, UDMA, TEGDMA BisGMA, BisEMA, UDMA, TEGDMA TCD-uretano BisGMA, BisEMA, TEGDMA

 Bonded area = [ · (R + r)]

h2 + (R − r)

2

where  = 3.1416; R larger cavity radius, r smaller cavity radius; h cavity high.

Considered by their respectvily manufacter as low-shrinkage composites.

2.2.

a

Ivoclar Vivadent, Schaan, Liechtenstein Saremco, Rohnacker, Switzerland 3M ESPE 3M ESPE Heraus Kulzer GmbH, Hanau, Alemanha Bisco, Schamburg, IL, EUA 0.04–0.2 ␮m 0.07–2.6 ␮m 75 nm–1.4 ␮m 0.19–3.3 ␮m 5 nm–20 ␮m 0.06 ␮m 46% 50% 57% 60% 64% 74% Heliomolar (HM) ELS (EL)a Filtek Supreme (SU) Filtek Z250 (FZ) Venus Diamond (VD)a Aelite LS Posterior (AE)a

Elastic Modulus (GPa) Post-gel Shrinkage (%) Organic matrix Manufacturer Average size of filler particles Filler content (vol.) Material (abbreviation)

Table 1 – Materials used in the present study, filler content in volume, average size of filler particle and manufactures. The data of post-gel shrinkage and elastic modulus according to Boaro [28].

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

Microleakage and marginal gap analysis

Bovine incisors (n = 15) had their buccal surfaces flattened with #400 grit sandpaper to provide an enamel surface large enough to make sure the cavity margins were entirely surrounded by enamel, and then received cylindrical cavities with 4-mm diameter and 1.5-mm depth (C-factor: 2.5, volume: 19 mm3 ), all of them with enamel margins. The restorative procedure was the same as described for the push-out test. Immediately after polymerization, the restorations were ground and polished with silicon carbide sandpaper (grits 600–4000) to remove composite excess and expose the restoration margins. After 24 h storage in distilled water at 37 ◦ C, the specimens were sonicated for cleaning the surface. Then, the restored surfaces were molded using an addition silicone (Express XT, light consistency, 3M ESPE) and the impressions were poured with epoxy resin (Buhler Epothin, Epoxicure Resin, Lake Bluff, IL, USA). After 9 h at 37 ◦ C, the replicas were separated from the molds, fixed in metal stubs and coated with gold for analysis in a scanning electron microscope (LEO, AEG-Zeiss, Germany) under 200× magnification. Ten specimens of each experimental group were randomly selected for gap analysis. Each specimen required between 38 and 48 images using 200× magnification to scan the entire perimeter of the restoration. Examples of debonded and gapfree margins are shown in Fig. 1. ImageJ software (National Institute of Health, Bethesda, USA) was used to measure the length of the debonded segments at the enamel-composite margins, as well as the entire perimeter of the restoration. The scale bar of the SEM images was used for calibration. The value

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

987

Fig. 1 – Examples of specimens showing interfacial gap (left) and without gap (right).

obtained in millimeters was converted to percentage based on the total perimeter of the interface for each specimen. After the silicone impression was obtained, each specimen was coated with nail polish, except in an area of 1 mm around the restoration. They were immersed in 50% AgNO3 for 2 h in the dark, followed by a period of 6 h immersion in developing solution (Kodak, São José dos Campos, SP, Brazil) under fluorescent light. After that, the specimens were sectioned with 0.3-mm diamond discs under water cooling (1000 Isomet, Buehler Ltd., Lake Bluff, IL, USA) twice perpendicularly through the center of the restoration. Since some of the tooth substance was lost during sectioning (due to the thickness of the diamond disc) eight surfaces (rather than four pairs of adjacent surfaces) were considered for microleakage evaluation. Images of each surface were digitized using a 60× magnifying stereomicroscope (model SZ61, Olympus Inc., Tokyo, Japan) equipped with a CCD camera (Q-Color 3, Olympus). The depth of penetration of the tracer was measured (in mm) using the ImageJ software, and both the average penetration of the eight surfaces and the maximum dye penetration were recorded. Additionally, the thickness of the enamel layer was also recorded.

2.3.

(Scotchbond Multipurpose Plus, bottle 3, 3M ESPE), light-cured with a radiant exposure of 12 J/cm2 (400 mW/cm2 × 30 s). The rods were attached to a universal testing machine (Instron – Fig. 2). Those with 13 mm were attached to the lower clamp and those with 28 mm to the upper clamp. The space between them was fixed at 1.5 mm (C-Factor: 1.3, volume: 19 mm3 ). The composites were inserted into this space and shaped as a cylinder following the perimeter of the rods. An extensometer was attached to the rods (model 2630-101, Instron) to monitor the specimen height and provide a feedback to the testing machine to move the actuator in order to keep the specimen height into a minimum range. The value registered by the load cell corresponded to the force necessary to counteract the polymerization shrinkage force and maintain the specimen’s initial height. The tip of the light guide (VIP Jr, Bisco) was positioned in contact with the polished surface of the 13 mm rod. The irradiance effectively reaching the composite was determined using a radiometer (model 100, Demetron Res. Corp., Orange, California, EUA) and duration of the exposure was adjusted to obtain a radiant exposure of 18 J/cm2 . Force development was monitored for 10 min from the beginning of photoactivation and the

Polymerization stress

Polymerization stress was measured under two compliance conditions, defined by the material used as bonding substrate for the composite: PMMA (“high compliance”) and glass (“low compliance”). Rods with 4 mm in diameter were sectioned in segments with 13 or 28 mm in length. For the 13 mm rods, one of the surfaces was polished with silicon carbide sandpaper (600–2000 grit) and felt disks with alumina paste (Alumina 3, ATM, Altenkirchen, Germany) in order to allow the highest light transmission possible during photoactivation. The opposite surface and both surfaces of the 28 mm rods were sandblasted with aluminum oxide (250 ␮m). For the PMMA rods, the sandblasted surfaces received a layer of methyl methacrylate monomer (JET Acrílico Auto Polimerizante, Artigos Odontológicos Clássico, São Paulo, Brasil), while the glass rods received a layer of silane (Ceramic Primer, 3M ESPE), followed by two layers of unfilled resin

Fig. 2 – Experimental set up of the polymerization stress test.

988

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

Table 2 – Means (standard deviation) for bond strength, microleakage, gap and polymerization stress (obtained using PMMA or glass as bonding substrate). In the same column, means followed by the same letter are statistically similar (one-way ANOVA/Kruskal–Wallis, p > 0.05). For the polymerization stress data, in the same row means followed by the same lower case letter are statistically similar (two-way Kruskal–Wallis/ANOVA, p > 0.05). Composite

Bond strength (MPa)

Marginal gap (%)

Microleakage (mm) Average

Filtek Supreme Aelite LS Posterior Filtek Z250 Venus Diamond Heliomolar ELS

B

5.1 (2.1) 4.7 (2.0)B 6.8 (2.7)AB 7.1 (2.1)AB 6.4 (1.6)AB 7.9 (3.2)A

A

47 (5) 42 (6)A 27 (4)B 13 (3)D 20 (5)C 21 (7)BC

maximum nominal stress was calculated by dividing the maximum force value recorded by the cross-section of the rods (n = 5).

2.4.

Statistical analysis

Polymerization stress data were analyzed using two-way ANOVA (composite and compliance) and Tukey test. Microleakage and marginal gap were analyzed using Kruskal–Wallis due to the lack of homocedasticity. In both tests, the pre-set global significance level was 5%. Pearson’s tests were used to verify the presence of statistically significant correlations between polymerization stress (in both substrates, glass and PMMA) and bond strength, microleakage or marginal gaps. In order to be considered statistically significant, the critical “r” value (Pearson’s correlation coefficient) was 0.811, according to the number of data pairs (six) and global significance level of 5% [28]. Regression analyses involving the same variables were also performed to determine the equations for the regression curves.

3.

Results

Means and standard deviations, as well as the statistical analysis for bond strength, microleakage (maximum and average), marginal gap and polymerization stress data are shown in Table 2. For bond strength, the only significant differences were between the ELS (7.9 MPa) and both Filtek Supreme (5.1 MPa) and Aelite (4.7 MPa). The percentage of marginal gap incidence ranged from 13% (Venus Diamond) to 47% (Filtek Supreme), and four statistical subsets were observed. Regarding microleakage, composites showed the same ordering for average and maximum values. However, there were four statistical subsets for average microleakage and only three subsets for maximum microleakage. Maximum microleakage varied between 0.71 and 1.34 mm (Heliomolar and Filtek Supreme, respectively). Average microleakage ranged between 0.35 and 0.89 mm (ELS and Filtek Supreme, respectively). For polymerization stress data, a significant interaction between composite and compliance level was observed (p < 0.001). On both compliance levels, ELS showed the lowest stress values (high compliance: 2.5 MPa; low compliance: 2.1 MPa) and Filtek Supreme, the highest stress (high compliance: 3.5 MPa; low compliance: 8.2 MPa). ELS, Heliomolar and Venus Diamond presented statistically similar stress values

Polymerization stress (MPa)

Maximum A

0.89 (0.18) 0.78 (0.18)AB 0.62 (0.20)BC 0.45 (0.12)CD 0.44 (0.12) CD 0.35 (0.14)D

High compliance

A

A,b

1.34 (0.21) 1.22 (0.28)A 1.10 (0.34)AB 0.85 (0.27)BC 0.71 (0.28) C 0.73 (0.38)C

3.5 (0.3) 3.4 (0.5)AB,b 2.9 (0.3)ABC,b 2.7 (0.4)BC,a 2.6 (0.6)BC,a 2.5 (0.3)C,a

Low compliance 8.2 (0.6)A,a 5.0 (0.8)B,a 5.1 (0.4)B,a 3.7 (0.3)C,a 3.3 (0.5) C,a 2.1 (0.1)D,a

on both compliance levels, while Filtek Supreme, Aelite LS and Filtek Z250 presented statistically higher stress values in the low compliance system. Table 3 presents the Pearson’s correlation coefficients (r) between polymerization stress (for both compliance levels) and dependent variables, namely, bond strength, microleakage and marginal gap. For stress data obtained in the low compliance system, statistically significant correlations were obtained with microleakage and marginal gap, but not with bond strength. Stress data obtained in the high compliance system presented statistically significant correlations with the three interfacial quality tests. Moreover, the r-values were consistently higher for correlations involving stress data from the low compliance system, compared to the high compliance system. The results of the regression analyses are shown in Fig. 3.

4.

Discussion

The tested composites ranked similarly for polymerization stress (on both substrates) and microleakage (maximum and average). For marginal gap, the ranking was also quite similar to those tests, except for the fact that Venus Diamond presented the lowest gap percentage and intermediate stress and microleakage. Bond strength, however, ranked the tested composites in a slightly different order. Based on materials’ post-gel shrinkage and elastic modulus values [10], in general, materials presenting lower post-gel shrinkage showed better interfacial integrity, regardless of their elastic modulus. Also, in broad terms, materials with low post-gel shrinkage presented low stress values for both compliance conditions.

Table 3 – Pearson correlation coefficient (r) for the analyses presented in this study. Considering the number of data pairs analyzed (six), r-values higher than 0.811 are statistically significant. Polymerization stress High compliance Bond strength Microleakage average maximum Marginal gap ∗

Low compliance

−0.910

−0.744*

0.993 0.942 0.932

0.934 0.889 0.817

Non-significant, p > 0.05.

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

Average microleakage (mm)

1,0 SU

SU 0,8

AE

y= 0,498x -0,860 r=0,993 0,6

AE

y=0,095x + 0,156 r=0,934

FZ FZ

VD HM

0,4

HM

A

VD

high compliance low compliance

EL EL 0,2 0

2

4

6

8

10

Polymerization Stress (MPa) 50

y= 29,33x - 56,956 SU r=0,932

AE

Gap (%)

40

30

FZ EL EL

20

B

AE

FZ

y= 5,20x + 4,591 r=0,817

HM

HM VD

high compliance low compliance

VD

10 0

SU

2

4

6

8

10

Polymerization Stress (MPa)

Bond Strength (MPa)

10

C y= -0.432x + 8.306 r= -0.744

EL EL

8

VD

VD

FZ 6

HM

HM

FZ

SU 4

SU AE

AE y= -2.61x + 13.931 r= -0.910

high compliance low compliance

2 0

2

4

6

8

10

Polymerization Stress (MPa) Fig. 3 – Regression analysis of polymerization stress for both compliance levels and average microleakage (A); marginal gaps (B); and bond strength (C).

However, as it will be discussed below, the influence of elastic modulus cannot be disregarded. Bond strength values reflect the complex interaction between the bonding substrate and the mechanical properties of the adhesive system and the composite and, not surprisingly, the method used to evaluate bond strength significantly influences the results [29]. The main aspect of the push-out test that differentiates it from tensile and shear tests in flat bonded interfaces is the confinement of the composite within the walls of the prepared cavity, more akin to the conditions found in the clinic. The bonded area in the push-out is larger

989

than those found with shear and tensile tests. Therefore, there is a higher probability of incorporating relatively large defects in the adhesive layer during the specimen preparation, which may result in lower bond strength values [30]. As described previously, the only statistically significant differences were found between ELS and Aelite LS/Filtek Supreme. This finding may be explained because these composites exhibit extreme values of post-gel shrinkage and modulus. As shown in Table 1, data from a previous study shows that ELS presented the lowest post-gel shrinkage (0.35%) associated with the lowest modulus (2.0 GPa) among all the tested composites [27]. On the other hand, Aelite LS presents the highest modulus (9.3 GPa) associated with a relatively high post-gel shrinkage (0.51%), while Filtek Supreme showed the highest shrinkage (0.64%) associated with an intermediate modulus (6.0 GPa). It is important to notice that the composites considered as “low shrinkage” by the respective manufacturers presented bond strength values statistically similar to the conventional composites. Composite restorations free of marginal gaps are very difficult to find clinically [31,32]. A study evaluating gap formation “in vivo” and “in vitro” for five different composites observed that gap incidence was always higher in vivo, and a significant correlation could be found between the data obtained in vivo and in vitro [32], which increases the relevance of laboratory evaluations. Venus Diamond showed the lowest percentage of marginal gaps, statistically similar to ELS and Heliomolar. This lower gap formation may be ascribed to the low post-gel shrinkage of these composites, and has already been reported for Venus Diamond [33,34]. Aelite LS presented post-gel shrinkage similar to Filtek Supreme and Filtek Z250 [27], which could explain the statistically similar percentage of gap formation for these composites. The majority of studies that evaluated microleakage used semi-quantitative methods (scores) or limited their analysis to the maximum tracer penetration in each specimen [35,36]. In the present study, the mean microleakage of each specimen was also evaluated in order to obtain a more comprehensive analysis of what occurs at the tooth/restoration interface. Maximum microleakage presented a lower ability to differentiate the materials; in other words, all materials presented relatively high dye penetration in at least one of the eight areas of the interface inspected. However, when the microleakage of the eight surfaces was averaged, more statistical subsets were found. These findings suggest that depending on the criteria used to quantify microleakage, it is possible to have different results for the same group of materials. Besides those limitations, microleakage is dependent on the type of teeth used (bovine or human), and even in the same tooth, regardless of the adhesive system, there are more microleakage in dentin than in enamel, due to the higher permeability of the former [37]. Though microleakage tests have been under severe scrutiny, in the present study a more comprehensive analysis was undertaken, trying to increase the consistency of the results. The average enamel thickness of the restored cavities was 0.55 ± 0.29 mm. Table 4 shows that enamel thickness was statistically similar for all groups, so the dye penetration was not influenced by this factor. As an average, for Venus Diamond, Heliomolar and ELS, most of the dye penetration was restricted

990

13 13 14 12 9 9 2 1 1 3 4 3 – 1 – – 1 3 – – – – 1 – – – – – – – 79 72 50 22 19 14

∗∗∗

∗

∗∗

Average enamel thickness for the entire sample (n = 720): 0.55 (0.29) mm. For each composite, a total of 120 surfaces was analyzed (n = 15, each specimen with 8 surfaces). There are no statistical differences for the column (ANOVA/one-way) or row (Student’s t test) for the enamel thickness.

41 48 70 98 101 106 0.40 (0.27) 0.55 (0.20) 0.88 (0.59) 0.44 (0.31) 0.62 (0.31) 0.55 (0.36) 0.55 (0.13) 0.47 (0.09) 0.67 (0.22) 0.42 (0.12) 0.55 (0.12) 0.40 (0.08) 0.70 0.74 1.40 0.79 1.00 0.98 0.15 0.31 0.17 0.13 0.27 0.13 0.47 (0.22) 0.51 (0.16) 0.80 (0.47) 0.43 (0.21) 0.60 (0.26) 0.50 (0.26) Filtek Supreme Aelite LS Posterior Filtek Z250 Venus Diamond Heliomolar ELS

8 7 6 5 4 or less Enamel and dentin Average

Minimum

Maximum

Dye penetration beyond DEJ***

Dye penetration restricted to enamel***

Only in enamel

Specimens presenting microleakage in (surfaces) Surfaces presenting microleakage** Enamel thickness (mm)* Composite

Table 4 – Means (standard deviation), minimum and maximum values for enamel thickness considering all the analyzed surfaces, and for those surfaces presenting microleakage reaching beyond the enamel-dentin junction (DEJ) or restricted to enamel. Also, the number of surfaces presenting microleakage restricted to enamel or both enamel and dentin are shown. Distribution of surfaces presenting microleakage for each specimen (tooth) is presented also.

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

to the enamel. As mentioned for bond strength, the low shrinkage composites showed microleakage values statistically similar to conventional materials. Aelite LS, for example, presented high microleakage probably due to its high modulus and shrinkage. Venus Diamond presented microleakage similar to Filtek Z250. It is possible to assume that, despite the low shrinkage of Venus Diamond, its intermediate elastic modulus resulted in a microleakage statistically similar to composites with greater shrinkage. ELS stayed on the lower microleakage subset along with Heliomolar and Venus Diamond. It must be pointed out that dentin and enamel represent two different behaviors in terms of compliance. It is well known that adhesion to enamel is more reliable compared to dentin, but enamel has a higher elastic modulus, and that increases stress concentration at the enamel-composite interface, as already has been demonstrated [38–40], and, consequently, the risk of gap formation and microleakage. Low shrinkage composites presented polymerization stress values statistically similar to conventional composites. Venus Diamond and ELS were statistically similar to Heliomolar and Filtek Z250. Venus Diamond contains a flexible monomer (TCD-urethane), whereas ELS has no diluent monomer. These features resulted in low elastic modulus, which associated with low post-gel shrinkage reduced the stress developed by these composites. It is interesting to notice that the relatively low modulus and shrinkage of Venus Diamond and ELS are not due to a lower degree of conversion. The degree of conversion of these composites is similar to Filtek Z250, while Heliomolar has a relatively lower degree of conversion. [41]. Polymerization stress is not a material property, and therefore, values vary depending on the testing system used [42,43]. The testing device, bonding substrate, specimen geometry and dimensions, and the adhesive system used to bond the composite to the substrate may significantly influence the obtained value. In this study, the stiffness of the testing system was changed only by varying the substrate. Stress values were higher with the stiffer substrate (glass) for Aelite LS, Filtek Supreme and Filtek Z250 but not for Venus Diamond, Heliomolar and ELS. In a previous study, it was observed that the lower the elastic modulus of the composite, the closer are the stress values obtained in different substrates [23]. That can be explained by the fact that in high compliance systems, composites with relatively high elastic modulus deform the bonding substrate in the radial direction, which results in a lower value registered by the load cell [43]. Coincidentally, except for Aelite LS, the other five materials chosen for this study show a direct relationship between shrinkage and elastic modulus (R2 = 0.72; y = 0.06x + 0.22). For example, the material presenting the highest elastic modulus (Filtek Supreme) also presented the highest post-gel shrinkage, and vice versa: the material presenting the lowest elastic modulus (ELS) presented the lowest post-gel shrinkage. This explains the similar ranking of the materials in terms of stress, regardless of the system compliance [43]. The similar ranking on both systems justify the correlations obtained between stress and microleakage or gap for both levels of compliance. However, bond strength showed the lowest correlation coefficient with stress obtained in PMMA and showed no correlation with the stress obtained in glass. One aspect

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

that differentiates the push-out specimen from the one used for microleakage and marginal gap evaluation is that in the former there was no enamel margins and, therefore, bonding occurred entirely on dentin, a more heterogeneous substrate. Besides, interfacial debonding is the result of the mechanical stress generated by an external load and, though influenced by polymerization stress, it is not a direct consequence of its effect. The findings of the present study suggest that when different compliance levels are compared, composite polymerization stress data generated in a high compliance system (PMMA) are more strongly related to the interfacial quality of bonded restorations, as the correlation coefficients were higher than those obtained with stress data obtained from a low compliance system. Though direct comparisons between mechanical tests and the clinical situation should be avoided, this finding is likely a reflection of the fact that the tooth is not a highly rigid substrate, demonstrating a high deformability in response to polymerization shrinkage [40,44]. In conclusion, for the group of materials evaluated, polymerization stress values obtained at both compliance levels can be related to interfacial quality in vitro. Nevertheless, stress data from the high compliance system showed higher correlation coefficients than data from the low compliance system for microleakage and marginal gap formation, besides a significant correlation with bond strength values.

Acknowledgement The authors would like to acknowledge the financial support provided by FAPESP (2008/54456-7).

references

[1] Hilton TJ. Can modern restorative procedures and materials reliably seal cavities? In vitro investigations. Part 1. Am J Dent 2002;15(3):198–210. [2] Brunthaler A, Konig F, Lucas T, Sperr W, Schedle A. Longevity of direct resin composite restorations in posterior teeth. Clin Oral Investig 2003;7(2):63–70. [3] Manhart J. Direct composite restorations in posterior region: a case history using a nanohybrid composite. Dent Today 2004;23(11):66, 68–70. [4] Ferracane JL, Mitchem JC. Relationship between composite contraction stress and leakage in Class V cavities. Am J Dent 2003;16(4):239–43. [5] Feilzer AJ, De Gee AJ, Davidson CL. Setting stress in composite resin in relation to configuration of the restoration. J Dent Res 1987;66(11):1636–9. [6] Watts DC, Marouf AS, Al-Hindi AM. Photo-polymerization shrinkage-stress kinetics in resin-composites: methods development. Dent Mater 2003;19(1):1–11. [7] Lu H, Stansbury JW, Bowman CN. Towards the elucidation of shrinkage stress development and relaxation in dental composites. Dent Mater 2004;20(10): 979–86. [8] Sakaguchi RL, Wiltbank BD, Shah NC. Critical configuration analysis of four methods for measuring polymerization shrinkage strain of composites. Dent Mater 2004;20(4):388–96.

991

[9] Alster D, Feilzer AJ, de Gee AJ, Davidson CL. Polymerization contraction stress in thin resin composite layers as a function of layer thickness. Dent Mater 1997;13(3):146–50. [10] Alster D, Venhoven BA, Feilzer AJ, Davidson CL. Influence of compliance of the substrate materials on polymerization contraction stress in thin resin composite layers. Biomaterials 1997;18(4):337–41. [11] Charton C, Colon P, Pla F. Shrinkage stress in light-cured composite resins: influence of material and photoactivation mode. Dent Mater 2007;23(8):911–20. [12] Charton C, Falk V, Marchal P, Pla F, Colon P. Influence of Tg, viscosity and chemical structure of monomers on shrinkage stress in light-cured dimethacrylate-based dental resins. Dent Mater 2007;23(11):1447–59. [13] Condon JR, Ferracane JL. Assessing the effect of composite formulation on polymerization stress. J Am Dent Assoc 2000;131(4):497–503. [14] Goncalves F, Pfeifer CC, Stansbury JW, Newman SM, Braga RR. Influence of matrix composition on polymerization stress development of experimental composites. Dent Mater 2010;26(7):697–703. [15] Braga RR, Ballester RY, Ferracane JL. Factors involved in the development of polymerization shrinkage stress in resin-composites: a systematic review. Dent Mater 2005;21(10):962–70. [16] Witzel MF, Ballester RY, Meira JB, Lima RG, Braga RR. Composite shrinkage stress as a function of specimen dimensions and compliance of the testing system. Dent Mater 2007;23(2):204–10. [17] Calheiros FC, Braga RR, Kawano Y, Ballester RY. Relationship between contraction stress and degree of conversion in restorative composites. Dent Mater 2004;20(10):939–46. [18] Goncalves F, Kawano Y, Braga RR. Contraction stress related to composite inorganic content. Dent Mater 2010;26(7):704–9. [19] Cunha LG, Alonso RC, Pfeifer CS, Correr-Sobrinho L, Ferracane JL, Sinhoreti MA. Modulated photoactivation methods: Influence on contraction stress, degree of conversion and push-out bond strength of composite restoratives. J Dent 2007;35(4):318–24. [20] Calheiros FC, Sadek FT, Braga RR, Cardoso PE. Polymerization contraction stress of low-shrinkage composites and its correlation with microleakage in class V restorations. J Dent 2004;32(5):407–12. [21] Lee SY, Park SH. Correlation between the amount of linear polymerization shrinkage and cuspal deflection. Oper Dent 2006;31(3):364–70. [22] Braga RR, Ferracane JL, Condon JR. Polymerization contraction stress in dual-cure cements and its effect on interfacial integrity of bonded inlays. J Dent 2002;30(7–8):333–40. [23] Boaro LC, Goncalves F, Braga RR. Influence of the bonding substrate in dental composite polymerization stress testing. Acta Biomater 2010;6(2):547–51. [24] Goncalves F, Pfeifer CS, Meira JB, Ballester RY, Lima RG, Braga RR. Polymerization stress of resin composites as a function of system compliance. Dent Mater 2008;24(5):645–52. [25] Fleming GJ, Hall DP, Shortall AC, Burke FJ. Cuspal movement and microleakage in premolar teeth restored with posterior filling materials of varying reported volumetric shrinkage values. J Dent 2005;33(2):139–46. [26] Palin WM, Fleming GJ, Nathwani H, Burke FJ, Randall RC. In vitro cuspal deflection and microleakage of maxillary premolars restored with novel low-shrink dental composites. Dent Mater 2005;21(4):324–35. [27] Boaro LC, Goncalves F, Guimaraes TC, Ferracane JL, Versluis A, Braga RR. Polymerization stress, shrinkage and elastic modulus of current low-shrinkage restorative composites. Dent Mater 2010;26(12):1144–50.

992

d e n t a l m a t e r i a l s 3 0 ( 2 0 1 4 ) 984–992

[28] Triola MF. Intriduc¸ão a Estatística. 9a edic¸ão LTC; 1999. [29] Cardoso PE, Braga RR, Carrilho MR. Evaluation of micro-tensile, shear and tensile tests determining the bond strength of three adhesive systems. Dent Mater 1998;14(6):394–8. [30] Pashley DH, Ciucchi B, Sano H, Carvalho RM, Russell CM. Bond strength versus dentine structure: a modelling approach. Arch Oral Biol 1995;40(12):1109–18. [31] Opdam NJ, Roeters FJ, Feilzer AJ, Verdonschot EH. Marginal integrity and postoperative sensitivity in Class 2 resin composite restorations in vivo. J Dent 1998;26(7): 555–62. [32] Hannig M, Friedrichs C. Comparative in vivo and in vitro investigation of interfacial bond variability. Oper Dent 2001;26(1):3–11. [33] Takahashi H, Finger WJ, Wegner K, Utterodt A, Komatsu M, Wostmann B, et al. Factors influencing marginal cavity adaptation of nanofiller containing resin composite restorations. Dent Mater 2010;26(12):1166–75. [34] Kurokawa R, Finger WJ, Hoffmann M, Endo T, Kanehira M, Komatsu M, et al. Interactions of self-etch adhesives with resin composites. J Dent 2007;35(12):923–9. [35] Ateyah NZ, Elhejazi AA. Shear bond strengths and microleakage of four types of dentin adhesive materials. J Contemp Dent Pract 2004;5(1):63–73. [36] Santini A, Ivanovic V, Ibbetson R, Milia E. Influence of cavity configuration on microleakage around Class V restorations

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

bonded with seven self-etching adhesives. J Esthet Restor Dent 2004;16(2):128–35 [discussion 136]. Heintze SD. Clinical relevance of tests on bond strength, microleakage and marginal adaptation. Dent Mater 2013;29(1):59–84. Ausiello P, Apicella A, Davidson CL, Rengo S. 3D-finite element analyses of cusp movements in a human upper premolar, restored with adhesive resin-based composites. J Biomech 2001;34(10):1269–77. Rodrigues FP, Silikas N, Watts DC, Ballester RY. Finite element analysis of bonded model Class I ‘restorations’ after shrinkage. Dent Mater 2012;28(2):123–32. Tantbirojn D, Versluis A, Pintado MR, DeLong R, Douglas WH. Tooth deformation patterns in molars after composite restoration. Dent Mater 2004;20(6):535–42. Boaro LC, Versluis A, Braga RR. Degree of conversion and post-gel shrinkage of low shrinkage composites. J Dent Res 2010;89(B). Tantbirojn D, Pfeifer CS, Braga RR, Versluis A. Do low-shrink composites reduce polymerization shrinkage effects? J Dent Res 2011;90(5):596–601. Meira JB, Braga RR, Ballester RY, Tanaka CB, Versluis A. Understanding contradictory data in contraction stress tests. J Dent Res 2011;90(3):365–70. Versluis A, Tantbirojn D, Pintado MR, DeLong R, Douglas WH. Residual shrinkage stress distributions in molars after composite restoration. Dent Mater 2004;20(6):554–64.