Push-out bond strength of CAD/CAM-ceramic luted to dentin with self-adhesive resin cements

Transcription

1 dental materials 26 (2010) available at journal homepage: Push-out bond strength of CAD/CAM-ceramic luted to dentin with self-adhesive resin cements Simon Flury, Adrian Lussi, Anne Peutzfeldt, Brigitte Zimmerli Department of Preventive, Restorative and Pediatric Dentistry, School of Dental Medicine, University of Bern, Freiburgstrasse 7, CH-3010 Bern, Switzerland article info abstract Article history: Received 30 March 2010 Received in revised form 4 May 2010 Accepted 4 May 2010 Keywords: Luting agents Thermocycling Storage Ceramic restoration Computer-aided design/manufacturing Objectives. This study evaluated the initial and the artificially aged push-out bond strength between ceramic and dentin produced by one of five resin cements. Methods. Two-hundred direct ceramic restorations (IPS Empress CAD) were luted to standardized Class I cavities in extracted human molars using one of four self-adhesive cements (SpeedCEM, RelyX Unicem Aplicap, SmartCem2 and icem) or a reference etch-and-rinse resin cement (Syntac/Variolink II) (n = 40/cement). Push-out bond strength (PBS) was measured (1) after 24 h water storage (non-aged group; n = 20/cement) or (2) after artificial ageing with 5000 thermal cycles followed by 6 months humid storage (aged group; n = 20/cement). Nonparametrical ANOVA and pairwise Wilcoxon rank-sum tests with Bonferroni Holm adjustment were applied for statistical analysis. The significance level was set at = In addition, failure mode and fracture pattern were analyzed by stereomicroscope and scanning electron microscopy. Results. Whereas no statistically significant effect of storage condition was found (p = 0.441), there was a significant effect of resin cement (p < ): RelyX Unicem showed significantly higher PBS than the other cements. Syntac/Variolink II showed significantly higher PBS than SmartCEM2 (p < 0.001). No significant differences were found between SpeedCEM, SmartCem2, and icem. The predominant failure mode was adhesive failure of cements at the dentin interface except for RelyX Unicem which in most cases showed cohesive failure in ceramic. Significance. The resin cements showed marked differences in push-out bond strength when used for luting ceramic restorations to dentin. Variolink II with the etch-and-rinse adhesive Syntac did not perform better than three of the four self-adhesive resin cements tested Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. 1. Introduction Ceramic restorations such as inlays and onlays allow a wide range of tooth reconstructions with good clinical success [1 4]. Furthermore, computer-aided design/manufacturing (CAD/CAM) systems such as CEREC simplify fabrication of ceramic reconstructions and are an effective and reliable restorative option [5 7]. Even restorations on teeth with highly reduced macroretention geometry and only little enamel left have shown acceptable clinical outcome [8]. However, challenges remain when working with ceramics. Apart from failures due to insufficient ceramic layer thickness, other main reasons for extensive or total failure of ceramic Corresponding author. Tel.: ; fax: address: simon.flury@zmk.unibe.ch (S. Flury) /$ see front matter 2010 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. doi: /j.dental

2 856 dental materials 26 (2010) restorations are luting defects or wear of the resin cement between the ceramic restoration and the tooth substance [9 12]. The resin cements differ according to the pretreatment of dental tissues prior to cementation and can mainly be divided into three subgroups: (1) conventional etch-andrinse resin cements (cements used after application of an etch-and-rinse adhesive including separate acid etching), (2) self-etch resin cements (cements used after application of a self-etch adhesive), and (3) self-adhesive resin cements ( selfadhering cements used without application of any adhesive system) [13,14]. Etch-and-rinse resin cements are often timeconsuming to use as well as sensitive to handling due to the numerous pretreatment steps involved. With self-adhesive cements, efforts have been made to simplify the luting process while aiming to maintain a reliable as well as durable bond to dental tissues, especially to dentin where obtaining a reliable bond is more challenging than to enamel because of the higher water content of dentin. Many studies have determined the bond strength of different resin cements by use of microtensile or shear bond strength methods [11,15 20]. However, extended sample manipulation such as repeated cutting could influence the bond and the small specimen size of these methods is prone to pretesting failure. There is no consensus so far as how to deal with pretesting failures and whether to include or to exclude these failures in statistical analysis [21]. In contrast, with the push-out bond strength method less specimen manipulation is needed prior to testing that could influence the bond or lead to pretesting failures. Besides, the push-out method allows a sample preparation that mimics clinical conditions. Despite these advantages, only limited literature exists describing the use of push-out forces for bond strength testing of ceramic restorations luted to dentin [22 24]. Moreover, information about the performance of self-adhesive cements after storage and artificial ageing is rather sparse [13]. Therefore, the aim of this study was to compare the pushout bond strength between ceramic and dentin using five different resin cements when measured 24 h after the luting process or when measured after artificial ageing consisting of thermocycling followed by 6 months of humid storage. Failure modes were classified stereomicroscopically and a qualitative analysis of the fracture sites was done by scanning electron microscopy (SEM). The following hypotheses were tested: (1) Artificial ageing results in a decrease of the initial bond strength. (2) The bond strengths of self-adhesive resin cements are significantly lower than that of the reference etch-and-rinse resin cement. 2. Materials and methods 2.1. Standardized ceramic restoration preparation The crown of a caries-free, extracted human molar was ground flat (Struers LaboPol-21/Struers Silicon Carbide paper (SiC) grit #320, diameter 200 mm, Ballerup, Denmark) until the entire surface was in coronal dentin. An elliptic occlusal cavity (diameter: 4 mm 5 mm, depth: 3 mm) was prepared with a 6 conical diamond bur (40 m grit; Intensiv FG A10, Grancia, Switzerland). The cavity was inspected under a microscope at 15 magnification (Leica ZOOM 2000, Leica, Buffalo, NY, USA) to ensure the absence of residual fissures, cracks, or partial exposure of the dental pulp. The tooth was then coated with IPS Contrast Spray Chairside (Ivoclar Vivadent AG, Schaan, Liechtenstein) and an optical impression was taken with the camera of the Sirona CEREC 3 acquisition center (Sirona, Bensheim, Germany). The shape of the future restoration was adjusted on screen with the CEREC 3D Software (Version 3.10, Sirona). With the CEREC 3 CAD/CAM milling unit (Sirona) twohundred standardized ceramic restorations were milled from IPS Empress CAD blocks (HT A3, size I8, LOT-Nr: M19049, Ivoclar Vivadent AG). The restorations were inspected under a magnifying glass and the milling stub was removed with a carbide bur and an OptraFine F finishing rubber point (Ivoclar Vivadent AG). The restorations were then sonicated for 3 min in ethanol (Telsonic TUC-150, Bronschhofen, Switzerland) and air-dried. The tooth-sided surfaces of the restorations were etched for 60 s with IPS Ceramic Etching (<5% hydrofluoric acid, LOT-Nr: M13217, Ivoclar Vivadent AG), cleaned with water spray, sonicated again for 1 min in ethanol and air-dried. Monobond S silane coupling agent (LOT-Nr: J27895, Ivoclar Vivadent AG) was applied for 60 s and airdried Standardized tooth cavity preparation The ground human molar with the elliptic occlusal cavity was cleaned from IPS Contrast Spray Chairside and a socket including a mounting for CELAY Plus (Mikrona Technologie AG, Spreitenbach, Switzerland & Ismaning, Germany) was waxed-up. The waxed-up tooth was embedded in Biotan Vest MG material (Schütz Dental GmbH, Rosbach, Germany) and the cast was effused with Biotan Titan 1 (LOT-Nr: , Schütz Dental GmbH) in a DOR-A-MATIC Argon- Vacuum/High-pressure apparatus (Labtech, Schütz Dental GmbH). The obtained titanium model of the tooth with the elliptic occlusal cavity was mounted in the CELAY Plus copygrinding device (Mikrona Technologie AG). A final amount of two-hundred extracted and sound human third molars was used in this study. The teeth were cleaned under tap water with a scaler and a hard toothbrush to remove calculus and debris, stored in 1% chloramine solution and kept at 5 C until needed. On each day prior to the luting process of one group (n = 20), 20 randomly selected teeth were ground flat as previously explained. With the CELAY Plus device, elliptic occlusal cavities were copy-drilled from the titanium tooth model into coronal dentin of the human molars with a 40 m grit diamond bur (CELAY ZY-54S, Mikrona Technologie AG) under water-cooling. The teeth were then rinsed under tap water and briefly air-dried. The standardized cavities were inspected under a microscope at 15 magnification (Leica ZOOM 2000). Teeth showing residual fissures or cracks, parts of cavity walls in enamel, or exposure of the pulp were discarded and replaced. The group of molars with standardized cavities was then stored in an incubator (Incu-

3 dental materials 26 (2010) Table 1 Resin cements used in this study (manufacturer information). Variolink II Ivoclar Vivadent AG, Schaan, Liechtenstein LOT-Nr: M03806 (Base)/L51132 (Catalyst, high viscosity) Paste/Paste Base (% weight) Etch-and-rinse adhesive resin cement Methacrylates 26.3% 22% Filler 73.4% 77.2% m (mean 0.7 m) Initiators, stabilizers and pigments 0.4% 0.9% Catalyst (% weight) SpeedCEM Ivoclar Vivadent AG, Schaan, Liechtenstein LOT-Nr: M32347 Paste/Paste (Automix) Universal dual-curing, self-adhesive resin cement Base (% weight) Methacrylates 23.3% 26% Filler 75% 2.2% m (mean: 5 m) Initiators, stabilizers and pigments 1.7% 0.9% Catalyst (% weight) RelyX Unicem Aplicap 3M ESPE, St. Paul, MN, USA LOT-Nr: Powder/Liquid (Capsule) Universal dual-curing, self-adhesive resin cement Powder (% weight) Methacrylates 25% Filler 70% 90% of filler <12.5 m Initiators, stabilizers and pigments total: 5% additives (powder & liquid) Liquid (% weight) SmartCem2 DENTSPLY Caulk, Milford, DE, USA LOT-Nr: Paste/Paste (Automix) Universal dual-curing, self-adhesive resin cement Base (% weight) Methacrylates n.a. n.a. Filler Total: 69% Glass (mean): 3.8 m/aerosil (mean): 16 nm Initiators, stabilizers and pigments n.a. n.a. Catalyst (% weight) icem Heraeus Kulzer, Hanau, Germany LOT-Nr: Paste/Paste (Automix) Universal dual-curing, self-adhesive resin cement Base (% weight) Methacrylates n.a. n.a. Filler Total: 49% n.a. (sub-micron and micron sized particles) Initiators, stabilizers and pigments n.a. n.a. n.a. = no further or detailed information of manufacturer available. Catalyst (% weight) bat, Melag, Berlin, Germany) for at least 2 h at 37 C and 100% humidity to assure intraoral temperature before the luting procedure Luting procedures and artificial ageing The self-adhesive cements used were SpeedCEM, RelyX Unicem Aplicap, SmartCem2 and icem. The etch-and-rinse resin cement Variolink II with Syntac Adhesive served as reference. The resin cements were stored according to manufacturers instructions. Cements that needed to be stored in the refrigerator were taken out at least 4 h before use. Detailed information about the resin cements is listed in Table 1. For a constant baseline temperature of the teeth before cementation and to mimic an intraoral condition, only one of the 20 prepared molars was taken out of the incubator at a time. Immediately after, the resin cement was applied in the tooth cavity and the pretreated ceramic restoration was inserted. A custom-made weight with a cone point was used to seat the restoration and ensured a constant force of 5 N. Excessive cement was removed with disposable soft brushes (Ivoclar Vivadent AG). The restoration was then light-cured with a LED light-curing unit (Bluephase Polywave, Ivoclar Vivadent AG) for 60 s in the High -power mode. Light power density was verified to be at least 1200 mw/cm 2 with a radiometer (Bluephase Meter, Ivoclar Vivadent AG) at the beginning and end of each

4 858 dental materials 26 (2010) Table 2 Luting procedures. Resin cement Treatment steps (according to manufacturer instructions) Time Syntac/Variolink II (n = 40) Total etching gel (37% phosphoric acid, LOT-Nr.: M30960) Water spray Syntac primer (LOT-Nr.: M34555) Syntac adhesive (LOT-Nr.: M19675) Heliobond (LOT-Nr.: M08772) Variolink II (base/catalyst 1:1) application 15 s (dentin only) >10 s + air dry 15 s + air dry 10 s + air dry (no light-curing) SpeedCEM (n = 40) Water spray >10 s + air dry SpeedCEM application RelyX Unicem Aplicap (n = 40) Water spray Capsule activation Capsule mixing (CapMix, 3M ESPE) RelyX Unicem Aplicap application >10 s + air dry >2 s 15 s SmartCem2 (n = 40) Water spray >10 s + air dry SmartCem2 application icem (n = 40) Water spray >10 s + air dry icem application luting procedure. The five luting procedures are described in Table 2. Twenty teeth per cement were produced for the group without artificial ageing and these were stored for 24 h in tap water at 37 C before determination of the push-out bond strength (non-aged group). Another 20 teeth of every cement underwent artificial ageing (aged group) immediately after the luting procedure. Artificial ageing consisted of 5000 thermal cycles (5 C/55 C, dwell time: 30 s, transfer time 5 s) followed by storage for 6 months at 37 C and 100% humidity in an incubator (Memmert UM 500, Schwabach, Germany) Push-out bond strength Before the push-out test, every tooth was completely embedded in self-curing acrylic resin (Paladur, Heraeus Kulzer GmbH, Hanau, Germany) using cylindrical molds. The coronal surfaces of the specimens were then polished with #500 and #1000 SiC paper (Struers) to remove remnants of Paladur or cement. The coronal dentin/restoration part of the embedded teeth was then marked with a felt pen, and discs of 2 mm ( mm) were cut using a diamond blade low-speed saw (Isomet, Lake Bluff, IL, USA). Discs were stored in tap water after their thickness had been measured with a digital caliper (Mitutoyo IP 65, Kawasaki, Japan) to calculate the bonding surface (BSU (mm 2 )): as the lateral area of the ceramic restorations represented a trapeziform surface, the bonding surface was calculated as A trapezoid (mm 2 ) = height (mm) center line (mm). The mean center line of 10 restorations was measured and set to be mm. Consequently, the formula was: BSU (mm 2 ) = thickness (mm) mm. For the push-out test, the embedded dentin/restoration discs were placed in a Zwick Z010 universal testing machine (Zwick GmbH & Co, Ulm, Germany) fitted with a custom-made jig (M. E. Mueller Institute, Bern, Switzerland) and loaded at a cross-head speed of 1.0 mm/min. Push-out force was applied on the restorations from the apical side because of their conicity. The maximum force (F max (N)) was recorded (testxpert software, V9.0, Zwick GmbH) and the push-out bond strength value (PBS (MPa)) was calculated according to the formula: PBS (MPa) = F max (N)/BSU (mm 2 ) Failure mode and SEM documentation Additionally, failure modes were analyzed visually with a stereomicroscope at 40 magnification (Leica ZOOM 2000) and classified into 4 categories: (1) cohesive failure in dentin (failure with dentin involvement), (2) adhesive failure of cement at the dentin interface, (3) adhesive failure of cement at the ceramic interface, (4) cohesive failure in ceramic (failure with ceramic involvement). For a qualitative documentation of fracture sites with SEM, three specimens per failure mode were mounted on aluminum stubs and sputter coated with gold/palladium for 100 s at 50 ma (Balzers SCD 050, Balzers, Liechtenstein). SEM was performed with a Stereoscan S360 scanning electron microscope at 20 kv (Cambridge Instruments, Cambridge, UK). Digital SEM micrographs (of 100 and 1000 magnification, respectively) were taken (Digital Image Processing System, version , point electronic GmbH, Halle, Germany) Statistical analysis The sample size used in the study was determined on the basis of a series of preliminary tests after the level of significance had been set at = 0.05 (NCSS/PASS 2005, NCSS, Kaysville, UT, USA). The PBS results were statistically analyzed with SAS (SAS Institute Inc., Cary, NC, USA). A nonparametrical ANOVA model for two fixed factors after Brunner and Munzel [25] was used to analyze PBS values with regard to any influence of resin cement and of storage condition. Pairwise Wilcoxon rank-sum tests with Bonferroni-Holm adjustment for multiple testing were used to analyze differences in PBS among the cements. The level of significance was set at = Results 3.1. Push-out bond strength The results of the PBS tests are shown in Table 3 (median, lower/1st and upper/3rd quartiles, and maxima/minima;

5 dental materials 26 (2010) Table 3 Results of the push-out bond strength (PBS) test. Resin cement Non-aged group PBS values (MPa) Aged group PBS values (MPa) (min) lq median uq (max) (min) lq median uq (max) mean values (SD) mean values (SD) Variolink II (3.34) (10.66) (4.54) (13.97) 7.23 (2.11) 8.21 (2.51) SpeedCEM (2.31) (15.01) (2.84) (13.57) 6.83 (3.19) 6.34 (2.96) RelyX Unicem Aplicap (5.30) (13.35) (5.24) (13.20) 9.47 (2.35) 9.21 (2.61) SmartCem2 (2.52) (10.18) (3.57) (17.48) 5.68 (2.48) 6.88 (4.23) icem (3.98) (10.16) (4.30) (12.09) 6.44 (1.70) 7.19 (2.17) SD = standard deviation; lq = lower/1st quartile, uq = upper/3rd quartile, min = minima, max = maxima. Fig. 1 Push-out bond strength values of the non-aged and aged group of each resin cement (median, lower/1st and upper/3rd quartiles, and maxima/minima). mean value and standard deviation) and in Fig. 1 for the nonaged and the aged group. The nonparametric ANOVA showed no statistically significant effect of storage condition (p = 0.441). No interaction effect of cement and storage condition was found (p = 0.636). On the other hand, the ANOVA showed a statistically significant effect of cement (p < ). RelyX Unicem Aplicap yielded significantly higher PBS values than did any other cement (Table 4). Variolink II showed significantly higher PBS values than SmartCem2 (p < 0.001) but no statistically significant differences compared to SpeedCEM or icem. Furthermore, no significant differences were found between SpeedCEM, SmartCem2, and icem (Table 4) Failure mode and SEM documentation In the aged as well as in the non-aged group, the predominant failure mode observed was adhesive failure of cement at Table 4 Results of the pairwise Wilcoxon rank-sum tests. Resin cement Variolink II SpeedCEM RelyX Unicem Aplicap SmartCem2 icem Variolink II * * SpeedCEM 1.21E 5 * RelyX Unicem Aplicap 3.71E 7 * 4.59E 6 * SmartCem p-values show significant differences after Bonferroni-Holm adjustment.

6 860 dental materials 26 (2010) Table 5 Distribution of failure modes per group and resin cement. Non-aged group (n = 20/cement) Aged group (n = 20/cement) Resin cement Specimens per failure mode Specimens per failure mode Variolink II SpeedCEM RelyX Unicem Aplicap SmartCem icem Failure modes 1: cohesive failure in dentin; 2: adhesive failure of cement at the dentin interface; 3: adhesive failure of cement at the ceramic interface; 4: cohesive failure in ceramic. failure in dentin (failure mode 1). Variolink II and SpeedCEM each presented one case of adhesive failure of the cement at the ceramic interface (failure mode 3) in the non-aged group, and RelyX Unicem Aplicap presented one case of failure mode 3 in the aged group. SmartCem2 presented more cohesive failures in ceramic (failure mode 4) in the aged group than in the non-aged group. SEM micrographs with representative fracture sites of all four failure modes are shown in Figs Discussion Fig. 2 Failure mode 1: fractures in dentin. the dentin interface (failure mode 2; Table 5) except for RelyX Unicem Aplicap which in most cases led to cohesive failure in ceramic (failure mode 4) followed by adhesive failure of cement at the dentin interface (failure mode 2) and cohesive Self-adhesive cements contain acidic monomers for tooth surface conditioning without further pretreatment and are reported to be less susceptible to moisture [26,27]. The dentin smear-layer is not completely removed by self-adhesive cements but modified. For the most commonly evaluated selfadhesive cement (RelyX Unicem Aplicap, 3M ESPE), it has been shown that no interface with a hybrid layer or deep resin tags is established [28,29]. Although no traditional etch-and-rinse adhesive interface can be seen with these cements, various studies have reported self-adhesive cements to show a reliable bond to dentin and good marginal adaptation to dentin or enamel [14,19,30,31]. However, self-adhesive cements are not Fig. 3 (a) Failure mode 2: resin cement (RC) on ceramic restoration (CI). (b) Failure mode 2: dentin interface free of resin cement.

7 dental materials 26 (2010) Fig. 4 Failure mode 3: resin cement (RC) on dentin (D). a homogeneous material group and the present study found significant differences in the bond to dentin between the cements investigated. Likewise, other studies have reported widely varying performances of self-adhesive cements not only regarding bond strength to dentin but also regarding water sorption and solubility [32]. Nevertheless, there is little knowledge about long-term performance of self-adhesive cements. Kitasako et al. showed that the bond strength of different resin cements decreased the most during the first 6 months of a 36 months storage [15]. Thus, in the present study half a year of storage in 100% humidity at 37 C was chosen in an effort to initiate a hydrolytic degeneration at the bonding interface and within the resin cement. To increase the artificial ageing process additionally, thermocycling was applied before storage to simulate stress due to thermal expansion and shrinkage. In literature, various effects of thermocycling have been reported: in a study of Brunzel et al., no statistically significant difference was observed between the bonding performance of RelyX Unicem after a soft thermocycling (37,500 cycles from 20 to 40 C) and Fig. 5 Failure mode 4: ceramic restoration partly fractured. a hard thermocycling (37,500 cycles from 5 to 55 C) [33]. On the other hand, Holderegger et al. found that a thermocycling procedure of 1500 cycles between 5 and 55 C influenced the bonding performance of all luting cements tested, however RelyX Unicem was least affected [34]. In the present study, thermocycling followed by 6 months storage had no statistically significant influence on the bonding performance (hypothesis 1 rejected). Especially when thermocycling entire teeth, one can assume that both hydrolytic degeneration and temperature differences at the luting interface are not as challenging as when microspecimens are thermocycled. In a future study, mechanical stress (e.g. in form of chewing simulation) instead of thermocycling and an enhanced storage time of 12 months could be considered to increase the artificial ageing process. With regard to the resin cements used, this study showed that Variolink II used with the etch-and-rinse adhesive system Syntac did not in general perform better than the self-adhesive cements on dentin: except for SmartCem2, the bond strength values of Variolink II were either equal to the self-adhesive cements or below (hypothesis 2 rejected). The performance of Variolink II compared to that of self-adhesive cements on dentin is controversially discussed in literature. In one study, Variolink II performed significantly better than RelyX Unicem, however a different adhesive system (Excite DSC) was used [17]. In another study, Variolink II with the Syntac adhesive system showed significantly higher bond strength values than RelyX Unicem and icem [35]. However, bond strength was determined with the shear bond method. On the other hand, there are various studies that confirm the present findings and in which Variolink II with Syntac yielded equal or significantly lower bond strength values than RelyX Unicem [14,36 38]. Statistically equal bond strengths on dentin have also been reported between self-etch resin cements and RelyX Unicem [14,28,35,36,38]. Clinically, however, resin cements are not only supposed to provide a reliable bond to dentin but also to preparation margins in enamel in order to reduce microleakage and thus marginal discoloration. On enamel, various studies have shown significantly lower bond strengths for self-adhesive than for etch-and-rinse cements [35,36,38], and regarding marginal quality, Frankenberger et al. concluded that the etchand-rinse technique provided the most reliable way for a good enamel margin adaptation [39]. In order to improve the bond of the self-adhesive resin cement RelyX Unicem to enamel, De Munck et al. recommended to selectively acid etch the enamel margins before application of the self-adhesive cement [28]. However, a recent clinical study has shown that selective acid etching of the enamel prior to application of RelyX Unicem had no significant influence on marginal integrity after 2 years [40]. When evaluating self-adhesive cements, RelyX Unicem has always yielded higher bond strength values than most of the other self-adhesive cements [13,14,20]. In the present study, RelyX Unicem Aplicaps (delivered as a liquid powder system) were used, whereas the other resin cements were paste paste systems. Whether the delivery form of RelyX Unicem Aplicap in any way contributed to the superior bonding performance of this cement is unclear. Future studies should investigate the stability of self-adhesive cements with regard to bonding

8 862 dental materials 26 (2010) effectiveness depending on delivery form, storage time, and storage condition. As confirmed recently, bond strength is also highly dependent on the test method as well as testing parameters applied [21]. Mostly, microtensile or shear bond tests have been applied [11,15 20]. However, the present study made use of a push-out test that had some advantages: (1) ceramic restorations were made with a common CAD/CAM system and cavities were drilled in human molars, mimicking a clinical situation, (2) no pretest failures occurred with this set-up which simplified the statistical analysis, (3) due to the bigger size of the push-out discs, correct positioning of the specimens in the universal testing machine was ensured, and (4) the failure mode could easily be classified. Except for RelyX Unicem Aplicap the predominant failure mode was adhesive failure of the cement at the dentin interface. Consequently, it can be assumed that the obtained PBS values in fact correlate with the bonding effectiveness of these resin cements to dentin. The superiority of RelyX Unicem Aplicap as regards PBS values on dentin was reflected in the failure mode as the predominant failure mode observed with this cement was cohesive failure in ceramic and not adhesive failure of the cement at the dentin interface. 5. Conclusions Within the limitations of the current in vitro study, the following conclusions may be drawn: The artificial ageing used in the present study had no significant effect on the push-out bond strength. On dentin, Variolink II with the etch-and-rinse adhesive system Syntac did not perform better than three of the four self-adhesive cements tested. When used for luting ceramic restorations to dentin, RelyX Unicem Aplicap yielded the highest push-out bond strength values and showed more cohesive failures in ceramic and dentin than the other resin cements. Conflicts of interest The authors declare no conflicts of interest, real or perceived, financial or nonfinancial. Acknowledgments The authors would like to thank 3M ESPE Switzerland, Dentsply DeTrey Germany, Heraeus Kulzer Switzerland and Ivoclar Vivadent Liechtenstein for providing the materials needed. Furthermore, we thank S. Hayoz and Prof. Dr. J. Hüsler, Institute of Mathematical Statistics and Actuarial Science, University of Bern for statistical advice. references [1] Hayashi M, Tsuchitani Y, Miura M, Takeshige F, Ebisu S. 6-Year clinical evaluation of fired ceramic inlays. Oper Dent 1998;23: [2] Hayashi M, Tsuchitani Y, Kawamura Y, Miura M, Takeshige F, Ebisu S. Eight-year clinical evaluation of fired ceramic inlays. Oper Dent 2000;25: [3] Molin MK, Karlsson SL. A randomized 5-year clinical evaluation of 3 ceramic inlay systems. Int J Prosthodont 2000;13: [4] Frankenberger R, Taschner M, Garcia-Godoy F, Petschelt A, Krämer N. Leucite-reinforced glass ceramic inlays and onlays after 12 years. J Adhes Dent 2008;10: [5] Hulterström AK, Bergman M. Polishing systems for dental ceramics. Acta Odontol Scand 1993;51: [6] Fasbinder DJ. Clinical performance of chairside CAD/CAM restorations. J Am Dent Assoc 2006;137(Suppl.):22S 31S. [7] Wittneben JG, Wright RF, Weber HP, Gallucci GO. A systematic review of the clinical performance of CAD/CAM single-tooth restorations. Int J Prosthodont 2009;22: [8] Bindl A, Richter B, Mörmann WH. Survival of ceramic computer-aided design/manufacturing crowns bonded to preparations with reduced macroretention geometry. Int J Prosthodont 2005;18: [9] Van Dijken JW, Höglund-Aberg C, Olofsson AL. Fired ceramic inlays: a 6-year follow up. J Dent 1998;26: [10] Van Dijken JW. Resin-modified glass ionomer cement and self-cured resin composite luted ceramic inlays. A 5-year clinical evaluation. Dent Mater 2003;19: [11] Escribano N, De La Macorra JC. Microtensile bond strength of self-adhesive luting cements to ceramic. J Adhes Dent 2006;8: [12] Manhart J, Chen H, Hamm G, Hickel R. Buonocore Memorial Lecture. Review of the clinical survival of direct and indirect restorations in posterior teeth of the permanent dentition. Oper Dent 2004;29: [13] Radovic I, Monticelli F, Goracci C, Vulicevic ZR, Ferrari M. Self-adhesive resin cements: a literature review. J Adhes Dent 2008;10: [14] Sarr M, Mine A, De Munck J, Cardoso MV, Kane AW, Vreven J, et al. Immediate bonding effectiveness of contemporary composite cements to dentin. Clin Oral Invest 2009 [Epub ahead of print]. [15] Kitasako Y, Burrow MF, Kathahira N, Nikaido T, Tagami J. Shear bond strengths of three resin cements to dentine over 3 years in vitro. J Dent 2001;29: [16] El Zohairy AA, De Gee AJ, Mohsen MM, Feilzer AJ. Microtensile bond strength testing of luting cements to prefabricated CAD/CAM ceramic and composite blocks. Dent Mater 2003;19: [17] Piwowarczyk A, Bender R, Ottl P, Lauer HC. Long-term bond between dual-polymerizing cementing agents and human hard dental tissue. Dent Mater 2007;23: [18] Cantoro A, Goracci C, Papacchini F, Mazztelli C, Fadda GM, Ferrari M. Effect of pre-cure temperature on the bonding potential of self-etch and self-adhesive resin cements. Dent Mater 2008;24: [19] D Arcangelo C, De Angelis F, D Amario M, Zazzeroni S, Ciampoli C, Caputi S. The influence of luting systems on the microtensile bond strength of dentin to indirect resin-based composite and ceramic restorations. Oper Dent 2009;34: [20] Viotti RG, Kasaz A, Pena CE, Alexandre RS, Arrais CA, Reis AF. Microtensile bond strength of new self-adhesive luting agents and conventional multistep systems. J Prosthet Dent 2009;102: [21] Scherrer S, Cesar PF, Swain MV. Direct comparison of the bond strength results of the different test methods: a critical literature review. Dent Mater 2010;26:e [22] Frankenberger R, Sindel J, Krämer N, Petschelt A. Dentin bond strength and marginal adaptation: direct composite resins vs ceramic inlays. Oper Dent 1999;24:

9 dental materials 26 (2010) [23] Kienanen P, Alander P, Lassila LV, Vallittu PK. Bonding of ceramic insert to a laboratory particle filler composite. Acta Odontol Scand 2005;63: [24] Cekic I, Ergun G, Uctasli S, Lasilla LVJ. In vitro evaluation of push-out bond strength of direct ceramic inlays to tooth surface with fibre reinforced composite at the interface. J Prosthet Dent 2007;97: [25] Brunner E, Munzel U. Nichtparametrische Datenanalyse (in German). Heidelberg: Springer; 2002, ISBN [26] Moszner N, Salz U, Zimmermann J. Chemical aspects of self-etching enamel dentin adhesives: a systematic review. Dent Mater 2005;21: [27] Mazzitelli C, Monticelli F, Osorio R, Casucci A, Toledano M, Ferrari M. Effect of simulated pulpal pressure on self-adhesive cements bonding to dentin. Dent Mater 2008;24: [28] De Munck J, Vargas M, Van Landuyt K, Hikita K, Lambrechts P, Van Meerbeek B. Bonding of an auto-adhesive luting material to enamel and dentin. Dent Mater 2004;20: [29] Monticelli F, Osorio R, Mazzitelli C, Ferrari M, Toledano M. Limited decalcification/diffusion of self-adhesive cements into dentin. J Dent Res 2008;87: [30] Mörmann W, Wolf D, Ender A, Bindl A, Göhring T, Attin T. Effect of two self-adhesive cements on marginal adaptation and strength of esthetic ceramic CAD/CAM molar crowns. J Prosthodont 2009;18: [31] Behr M, Hansmann M, Rosentritt M, Handel G. Marginal adaptation of three self-adhesive resin cements vs. a well-tried adhesive luting agent. Clin Oral Invest 2009;13: [32] Vrochari AD, Eliades G, hellwig E, Wrbas KT. Water sorption and solubility of four self-adhesive resin luting agents. J Adhes Dent 2010 [Epub ahead of print]. [33] Brunzel S, Yang B, Wolfart S, Kern M. Tensile bond strength of a so-called self-adhesive luting resin cement to dentin. J Adhes Dent 2009 [Epub ahead of print]. [34] Holderegger C, Sailer I, Schuhmacher C, Schläpfer R, Hämmerle C, Fischer J. Shear bond strength of resin cements to human dentin. Dent Mater 2008;24: [35] Lührs AK, Guhr S, Günay H, Geurtsen W. Shear bond strength of self-adhesive resins compared to resin cements with etch and rinse adhesives to enamel and dentin in vitro. Clin Oral Invest 2010;14: [36] Hikita K, Van Meerbeek B, De Munck J, Ikeda T, Van Landuyt K, Maida T, et al. Bonding effectiveness of adhesive luting agents to enamel and dentin. Dent Mater 2007;23: [37] Bitter K, Paris S, Pfuertner C, Neumann K, Kielbassa AM. Morphological and bond strength evaluation of different resin cements to root dentin. Eur J Oral Sci 2009;117: [38] Abo-Hamar SE, Hiller KA, Jung H, Federlin M, Friedl KH, Schmalz G. Bond strength of a new universal self-adhesive resin luting cement to dentin and enamel. Clin Oral Invest 2005;9: [39] Frankenberger R, Lohbauer U, Schaible RB, Nikolaenko SA, Naumann M. Luting of ceramic inlays in vitro: marginal quality of self-etch and etch-and-rinse adhesives versus self-etch cements. Dent Mater 2008;24: [40] Peumans M, De Munck J, Van Landuyt K, Poitevin A, Lambrechts A, Van Meerbeek B. Two-year clinical evaluation of a self-adhesive luting agent for ceramic inlays. J Adhes Dent 2009 [Epub ahead of print].