1

Introduction

Intra-oral scanning in combination with CAD/CAM fabrication allows for an innovative work flow utilizing the potential of different CAD/CAM materials. CAM-machinable materials are available as blocks fabricated from composite, resin-infiltrated ceramic, or different ceramics (feldspar, lithium disilicate, zirconia-reinforced lithium silicate (ZLS), zirconia). In most cases, chairside fabrication of these materials is optimized for small restorations such as single crowns, veneers, onlays, or smaller anterior fixed partial dentures. Composite materials, which do not require any additional treatment after milling such as sintering, crystallization, or glazing feature a high potential for application as cost- and time-effective chairside restorations. The currently available materials differ in composition and mechanical properties , in-vitro performance and strength , fracture toughness , machinability , or grinding effects . For example, composites and ceramics show strongly different moduli of elasticity or coefficients of thermal expansion .

In 2015, the application of a CAD/CAM resin-based composite (Lava Ultimate, 3 M Oral Care, USA) was astonishingly restricted to veneers, inlays, and onlays, as for crowns a high number of debondings ranging around 10% was observed . As a result of the limited clinical experience and the lack of prospective studies, it is possible that similar problems might occur with other novel resin-based materials, too.

While the reasons for the increased number of crown debondings are still unclear, bonding appears to be not the only reason for high failure rates, as the material is still available and can be applied in other indications than crowns. Other potential reasons for the clinically observed failures might include water uptake, insufficient polymerization of the adhesive, inadequate crown pretreatment or cementation, insufficient preparation or fitting or combinations. The mentioned composite seemed to be more prone to water uptake than a hybrid material and showed hardness reduction during water storage . Bonding is essential for the success of these materials ; however, resin-based materials feature differences in light transmission , which might limit the activation of adhesive and/or cement. With regard to this aspect, transmission appeared to be sufficient for clinical application up to a restoration thickness of 2 mm . Moreover, bond strength is dependent on the mode of curing . Additional factors impacting bonding of CAD/CAM composites include pre-treatment and the type of cementation . With regard to these aspects, the application of silane may not regularly support bonding in all cases , and adhesive bonding appears to be more effective than self-adhesive cementation .

The occlusal load during chewing in combination with a high elasticity (E-modulus composite: 13 GPa, lithium disilicate ceramic: 100 GPa) of the composite crown may cause pumping or deformation as well as spreading of the margins, which might finally result in debonding. Occlusal cement gaps, which differ between individual composites, seem to be larger than axial gaps . These negative effects may be crucial for an effective cementation of composite restorations, because resin-based materials show a higher stress distribution at the central bonding interface in comparison to ceramics . Insufficient preparation or fit may further enhance these effects.

Although already in clinical use, there are only little scientific information and even less clinical data which support the long-term applicability of CAD/CAM composites. Laboratory tests such as thermal cycling and mechanical loading may allow a first prediction of the long-term mechanical performance of the materials, their resistance against hydrolytic effects, or the influences of preparation design and fitting. The methods employed in laboratory approaches can enhance fatigue or debonding and provide detailed information on probable failures. However, even in cases without any visible failures during simulation, aging and deteriorating effects might occur, which reduce strength and fracture resistance. In these cases, a subsequent static fracture test may help to identify weak points or initial debondings of the restoration.

The hypothesis of this investigation was that CAD/CAM-fabricated posterior composite crowns show different in-vitro performance and fracture resistance as a function of (a) preparation design, (b) fitting parameters from milling, and (c) type of cementation.

2

Materials and methods

Extracted human molars (n = 8 per group) were prepared with two different preparation designs. The retentive design was based on ceramic guidelines with a circular and occlusal anatomical reduction of about 1.5 mm (retentive preparation (R): height ∼8 mm; angle ∼8°). The non-retentive design was characterized by reduced teeth height and increased preparation angle (no retention (NR): height ∼4 mm; angle ∼15°). In all cases, the finishing line resulted in an approx. 1 mm deep circular shoulder with rounded inner angles at an isogingival height of the tooth cervix ( Fig. 1 ). All teeth were positioned in resin blocks (Palapress Vario, Heraeus Kulzer, D) and resilience of the human periodontium was simulated by coating the roots of the teeth with a 1 mm polyether layer (Impregum, 3 M Oral Care, USA). For achieving a constant layer, the roots were dipped in a wax bath, which was replaced by polyether in a second fabrication process .

Micro-CT picture of a separated crown; cement (dark) is fixed to the crown (light) (courtesy of Regensburg Center of Biomedical Engineering (RCBE)).

Prepared teeth were digitalized (Cerec Omnicam, Sirona, D) and full-contour molar crowns were milled (Cerec, MC XL, Sirona, D). The circular wall thickness was 1.5 mm and occlusal dimensions varied according to the preparation design between 1.5 mm (R) and 5.5 mm (NR). Fitting parameters were varied between optimal fit (OF: 100 μm) and reduced fit (RF: 250 μm). A total of 112 molar crowns were milled from two resin-based composite CAD/CAM materials and one ceramic reference (n = 8 per material per group, Table 1 ). The inner surfaces of the crowns were treated as recommended by the manufacturers ( Table 1 ). For self-adhesive cementation (SE), Rely X Unicem 2 Automix (3 M Oral Care) was used, and adhesive bonding (A) was performed with Bifix QM (Voco, D; Elipar Trilight, 3 M ESPE, USA; 5 × 40 s) after conditioning of tooth tissues (Futurabond U, Voco, D; total etch; 10 s).

For investigating the influence of tooth structure, two identical crown series were fabricated on implant abutment analogues (n = 2 × 8; Straumann, D; titanium grade IV; implant diameter 4.1 mm and length 12 mm; abutment length 4 mm, 8°), which were vertically positioned in resin blocks (Palapress Vario, Heraeus Kulzer) in order to simulate the posterior implant situation replacing tooth 46 with a non-retentive abutment. Crowns were luted with self-adhesively (SE) using Rely X Unicem 2 Automix (3 M Oral Care, USA) after tribochemical activation (Rocatec, 3 M Oral Care, USA) of the abutments.

Restored restorations were stored in distilled water for 90 days (37 °C). Thermal cycling and mechanical loading (TC: 2 × 3000 cycles between 5 °C/55 °C, dist. water, ML: 50 N for 1.2 × 10 ⁶ cycles; f = 1.6 Hz; mouth opening: 2 mm; chewing simulator EGO, Regensburg, Germany) with online failure-control was performed to stimulate and control fatigue failures. Steatite balls with a diameter of 12 mm (CeramTec, D) served as standardized antagonists and were positioned in a three point occlusal contact situation. Failures were documented; failed specimens were excluded from the simulation process. Chewing simulation parameters were based on literature data regarding zirconia and ceramic restorations, assuming that the chewing simulation performed simulates five years of oral service . Restorations which failed during TCML were investigated in detail using scanning electron microscopy (SEM Quanta, Philips, NL). All restorations that survived TCML were loaded to fracture (1446, Zwick, v = 1 mm/min). Analogously to chewing simulation, the load was occlusally applied with a steel sphere (d = 12 mm). A tin foil (0.25 mm, Dentaurum, D) between crown and sphere prevented force peaks. All systems were optically examined after fracture testing and the failure mode was documented.

Calculations and statistical analyses were performed using SPSS 23.0 for Windows (SPSS Inc., Chicago, IL, USA). Power calculation (G*Power 3.1.3, Kiel, D) provided an estimated power of >90% using eight specimens per group. Means and standard deviations were calculated and analyzed using one-way analysis of variance (ANOVA) guided by Bonferroni-test for post-hoc analyses where appropriate. The level of significance (α) was set to 0.05. Failed or debonded crowns were excluded from evaluation.

2

Materials and methods

Extracted human molars (n = 8 per group) were prepared with two different preparation designs. The retentive design was based on ceramic guidelines with a circular and occlusal anatomical reduction of about 1.5 mm (retentive preparation (R): height ∼8 mm; angle ∼8°). The non-retentive design was characterized by reduced teeth height and increased preparation angle (no retention (NR): height ∼4 mm; angle ∼15°). In all cases, the finishing line resulted in an approx. 1 mm deep circular shoulder with rounded inner angles at an isogingival height of the tooth cervix ( Fig. 1 ). All teeth were positioned in resin blocks (Palapress Vario, Heraeus Kulzer, D) and resilience of the human periodontium was simulated by coating the roots of the teeth with a 1 mm polyether layer (Impregum, 3 M Oral Care, USA). For achieving a constant layer, the roots were dipped in a wax bath, which was replaced by polyether in a second fabrication process .

Micro-CT picture of a separated crown; cement (dark) is fixed to the crown (light) (courtesy of Regensburg Center of Biomedical Engineering (RCBE)).

Prepared teeth were digitalized (Cerec Omnicam, Sirona, D) and full-contour molar crowns were milled (Cerec, MC XL, Sirona, D). The circular wall thickness was 1.5 mm and occlusal dimensions varied according to the preparation design between 1.5 mm (R) and 5.5 mm (NR). Fitting parameters were varied between optimal fit (OF: 100 μm) and reduced fit (RF: 250 μm). A total of 112 molar crowns were milled from two resin-based composite CAD/CAM materials and one ceramic reference (n = 8 per material per group, Table 1 ). The inner surfaces of the crowns were treated as recommended by the manufacturers ( Table 1 ). For self-adhesive cementation (SE), Rely X Unicem 2 Automix (3 M Oral Care) was used, and adhesive bonding (A) was performed with Bifix QM (Voco, D; Elipar Trilight, 3 M ESPE, USA; 5 × 40 s) after conditioning of tooth tissues (Futurabond U, Voco, D; total etch; 10 s).

For investigating the influence of tooth structure, two identical crown series were fabricated on implant abutment analogues (n = 2 × 8; Straumann, D; titanium grade IV; implant diameter 4.1 mm and length 12 mm; abutment length 4 mm, 8°), which were vertically positioned in resin blocks (Palapress Vario, Heraeus Kulzer) in order to simulate the posterior implant situation replacing tooth 46 with a non-retentive abutment. Crowns were luted with self-adhesively (SE) using Rely X Unicem 2 Automix (3 M Oral Care, USA) after tribochemical activation (Rocatec, 3 M Oral Care, USA) of the abutments.

Restored restorations were stored in distilled water for 90 days (37 °C). Thermal cycling and mechanical loading (TC: 2 × 3000 cycles between 5 °C/55 °C, dist. water, ML: 50 N for 1.2 × 10 ⁶ cycles; f = 1.6 Hz; mouth opening: 2 mm; chewing simulator EGO, Regensburg, Germany) with online failure-control was performed to stimulate and control fatigue failures. Steatite balls with a diameter of 12 mm (CeramTec, D) served as standardized antagonists and were positioned in a three point occlusal contact situation. Failures were documented; failed specimens were excluded from the simulation process. Chewing simulation parameters were based on literature data regarding zirconia and ceramic restorations, assuming that the chewing simulation performed simulates five years of oral service . Restorations which failed during TCML were investigated in detail using scanning electron microscopy (SEM Quanta, Philips, NL). All restorations that survived TCML were loaded to fracture (1446, Zwick, v = 1 mm/min). Analogously to chewing simulation, the load was occlusally applied with a steel sphere (d = 12 mm). A tin foil (0.25 mm, Dentaurum, D) between crown and sphere prevented force peaks. All systems were optically examined after fracture testing and the failure mode was documented.

Calculations and statistical analyses were performed using SPSS 23.0 for Windows (SPSS Inc., Chicago, IL, USA). Power calculation (G*Power 3.1.3, Kiel, D) provided an estimated power of >90% using eight specimens per group. Means and standard deviations were calculated and analyzed using one-way analysis of variance (ANOVA) guided by Bonferroni-test for post-hoc analyses where appropriate. The level of significance (α) was set to 0.05. Failed or debonded crowns were excluded from evaluation.