The materials used for coronal restorations in this study are presented in Table (1). Sample size was calculated using G Power 3.1.9.7., based on the previous study by Hitz et al. [11]. , as reference. According to this study, the accepted sample size was 8 per group, when the mean and standard deviation of group 1 was (90.3 ± 2.6), while the mean and standard deviation of group 2 were (80.4 ± 9) with 1.23 effect size. Total sample size was increased to 10 to compensate for 20% drop out when the power was 80% & type I error probability was 0.05.

Table 1 Materials used in coronal restorations and their compositions:

This study proposal was reviewed and approved by the Institutional Review Board of the Faculty of Oral and Dental Medicine, Ahram Candian University (IRB00012891), Egypt, on 25/6/2023. With approval number IRB00012891 ≠ 71. A total of thirty sound human maxillary first premolars extracted for orthodontic purposes were obtained from the Department of Oral and Maxillofacial Surgery at the Faculty of Dentistry, Ahram Canadian University. Teeth were examined under magnification loups 4.0X and LED headlight (Univet, Italy) for caries, abrasions, cracks or fractures. Teeth with any of these defects were excluded. Chosen teeth were measured for length of 15 + 1.0 mm and similar buccolingual, mesiodistal, and occluso-gingival dimensions using digital caliper with 0.01 mm accuracy (Mitutoyo IP 65, Kawasaki, Japan). Debris were cleaned using ultrasonic scaler, followed by polishing with a rubber cup and pumice. Immersion in 0.1% thymol solution at room temperature was done for 1 day, and finally, teeth were stored at room temperature in deionized distilled water container until the experiment started [4, 5].

Endodontic treatment

Conservative access cavity preparation was done using endo-access round diamond bur (856; Intensiv SA, Switzerland). Following hybrid method, a low-speed contra-angle hand-piece (W&H Dentalwerk Bürmoos GmbH, Austria), was used for a step-down procedure using gates Glidden burs for the first 3 mm. A digital X-ray was used to assess the working length after the insertion of NiTi files. (Mani NiTi K File) A No.40 master apical file was used and a 1 mm steps step-back procedure was performed until file No.60. Sodium hypochlorite (5%, sodium hypochlorite solution) was used after each file to rinse the canal, and paper points are used to dry the canal. The obturation process was done using No.40 gutta-percha points, accessory points (Meta Biomed), and sealer (AD Seal Meta Biomed). After removal of excess gutta-percha, the access cavity was cleaned with alcohol [4].

Sample grouping

30 teeth were randomly divided into three equal experimental groups according to the coronal restoration after endodontic treatment:

Group R: Restored with non-cuspal coverage direct restoration; polyethylene fibers (ribbond), glass fibers reinforced resin composite (ever-x posterior) and a final layer of nano-hybrid composite (n = 10).

Group O: Restored with cuspal coverage, indirect lithium disilicate ceramic overlay (n = 10).

Group C: Restored with fiber-post, nano-hybrid resin composite restoration and lithium disilicate full coverage ceramic crown (n = 10).

For periodontal simulation, each tooth was embedded in molten wax (CAVEX, CAVEX dental, Netherland), 2 mm below the cementoenamel junction (CEJ). After complete wax solidification, teeth placed in a custom-made plastic mold (10 mm radius and 15 mm depth) filled with self-curing acrylic resin (Acrostone; Acrostone dental, Cairo, Egypt) simulating alveolar bone. After initial polymerization, wax remnants were cleaned out from teeth and acrylic, and light-body polyvinyl siloxane impression material (Panasil® Initial contact; Kettenbach GmbH & Co) was injected into the mold. Teeth were re-inserted parallel to its long-axis into the acrylic socket under constant finger pressure to extrude the excess silicon material and create a uniform silicon layer (0.2–0.3 mm) that simulate the periodontal ligament.

For both group R and group O, a thin layer of universal adhesive (All-Bond Universal. BISCO Inc, USA) was applied to the walls of pulp chamber, and light polymerized for 10 s with light curing unit (Elipar™, 3 M ESPE, USA) according to manufacturer instructions. The pulp chamber was then filled with flowable composite (Filtek ™ Flow, 3 M ™ ESPE, USA).

To ensure proper standardization of teeth preparation, biocopy mode on CEREC 3D software (version 4.5, Sirona Dental Systems GmbH, Bensheim, Germany) was used to obtain a base scan of all teeth before preparation. Samples were scanned using Omnicam scanner and base images were obtained and stored on CEREC 4.5 software to allow for superimposition of the biocopy and the post-preparation scans, so, standardized cavity preparation could be verified. Moreover, polyvinyl siloxane silicon index (Elite HD+, Zhermack-Germany) was taken for teeth before preparation for verification and for use as occlusal stamp with teeth assigned to direct restoration group (n = 10).

All endodontic procedures and teeth preparation steps were performed by the same operator, under magnification loups 4.0X. Moreover, all indirect restorations laboratory fabrication steps were done by the same laboratory technician.

Standardized MOD cavity was prepared for all teeth with mesial and distal gingival cavo-surface margins 1 mm above CEJ. Intra-coronal cavity of 4 mm depth was prepared using short parallel diamond bur (6836KR, Komet, Schaumburg, USA) mounted in high-speed handpiece (Synea WK-900 LT, W&H Dentalwerk Bürmoos GmbH, Austria) under copious amount of water coolant. The buccal and lingual residual wall thickness were adjusted to 1.5 ± 0.2 mm at the height of contour, and the bucco-lingual width of the proximal box was adjusted to 1/3 of the total bucco-lingual tooth dimensions. Internal line-angles were smoothed and all unsupported enamel was removed with subsequent finishing diamond bur (8836KR, Komet, Schaumburg, USA). All prepared cavities were verified for accurate dimensions with digital caliper.

Following this, three different adhesive coronal preparations were carried out for the three study groups (n = 10 each), Fig. (1):

Group R (direct restoration; ribbond + EverX resin composite)

Universal adhesive was applied to the cavity walls and cured, followed by the application of a thin layer of flowable resin composite. A thin strip of Ribbond Ultra (~ 5 mm long, 2 mm wide) was prepared and saturated with MDP containing bonding agent (Gluma® Bond Universal, Kulzer GmbH, Hanau, Germany) and placed in flowable resin in a bucco-lingual direction not touching the enamel margins. After light-curing for 20 s, the remaining cavity space was restored with fiber-reinforced resin composite EverX Posterior (GC Europe, Leuven, Belgium). The last layer with nano-hybrid resin composite (Filtek Z250-xt 3 M ™ ESPE, USA) was applied, then, finishing and polishing were done using discs (Soflex; 3 M ESPE, Germany) [12].

Group O (CAD/CAM- indirect overlay restoration)

Bonding was done to the cavity walls similar to group R, then, the MOD cavity except for the gingival seat (1 mm thickness) was restored with nano-hybrid composite applied in consecutive horizontal layers of maximum 3 mm thickness and light cured for 20 s each as per manufacturer recommendations.

Indirect adhesive overlay preparation was done following morphology driven preparation design (MDPD), with hollow chamfer margin by Veneziani [13]. First, proximal box was adjusted to rounded angles with 8°divergence, 1.5 mm width, 1 mm depth and located 1 mm above the cementoenamel junction using flat-end diamond bur (8845KR, Komet, Schaumburg, USA). Occlusal reduction was done by creating depth orientation grooves using depth-cut bur (DM20, Komet, Schaumburg, USA), followed by OccluShaper diamond bur (KP6370, Komet, Schaumburg, USA) to obtain uniform anatomical occlusal reduction of 1.5 mm, then fine diamond OccluShaper bur (8370, Komet, Schaumburg, USA) was used for occlusal surface finishing. Hollow chamfer margin of 0.8 mm (± 0.2) thickness was prepared using tapered chamfer diamond bur (6856-014, Komet, Schaumburg, USA) and finished with fine grit diamond bur (8856-014, Komet, Schaumburg, USA). All preparation corners were finally rounded with spitz pointed Arkansas white abrasive (307, Komet, Schaumburg, USA) and polisher (9436 M, Komet, Schaumburg, USA).

Group C: (Fiber post, composite restoration, and lithium disilicate full coverage ceramic crown)

For cylinder-tapered size 2 glass fiber post (FiberKleer™ 4X, Pentron, USA), a 10 mm deep preparation was made in the palatal canal using the corresponding, leaving 5 mm of gutta-percha as apical seal. Universal adhesive was applied to the cleaned and dried post space for 20 s then air dried. Dual cured adhesive resin cement (Duo-link Universal ™, Bisco Inc, Schaumburg, USA), was introduced into the post space with lentulo spiral rotary instrument (Dentsply Sirona, Switzerland), followed by immediate insertion of fiber post under finger pressure for 15 s. After complete removal of excess cement, light polymerization was done for 40 s. Cavities were bonded and restored with nano-hybrid composite as in group O [14].

For full coverage crown preparation, a standardized planner occlusal reduction of 2 mm was achieved. Axial reduction with heavy chamfer margin 1 ± 0.5 mm above cementoenamel junction was done with tapered chamfer diamond bur and finished with similar size fine grit diamond bur. All preparation corners were finally rounded with spitz pointed Arkansas white abrasive [15]. Preparations for all indirect adhesive restorations in both Group C and O were checked by digital caliper and digitally verified using CEREC software via PrepCheck feature (version 4.5, Sirona Dental Systems GmbH, Bensheim, Germany).

Fig. 1

Schematic representation of the three study groups, restored with different approaches: A: Ribbond fibers + EverX posterior composite (Group R), B: Ceramic overlay (Group O), and C: Fiber post + composite build-up + ceramic crown (Group C)

Digital workflow, fabrication of indirect restorations and final cementation

Omnicam intraoral scanner (Dentsply-Sirona, Bensheim, Germany) was used for scanning of the prepared teeth from the two indirect restorations groups (Group O and Group C). Restorations were designed using CEREC 3D software (version 4.5, Dentsply-Sirona, Bensheim, Germany). To standardize the design with similar occlusal morphology and average thickness of 1.5 mm at the cusp tip, and 1 mm fissure thickness, biogeneric reference tool in CEREC software was utilized. It allows designing all restorations from a previously stored dentoform maxillary first premolar scan. Only position tool was used for restoration adjustment to avoid any alteration of restoration’s original proposal. Cement space of 80 μm to allow precise restoration seating without marginal discrepancy [16, 17].

All restorations were milled from lithium disilicate blocks (IPS e.max CAD, Ivoclar Vivadent, Schaan, Liechtenstein) using MCXL 4-axis wet milling and grinding machine (Dentsply Sirona, Bensheim, Germany). After checking of thickness with digital caliper, crystallization and glazing were done in ceramic furnace Programat P310 (Ivoclar Vivadent Inc., New York, USA) following the recommended manufacturer parameters.

Adhesive cementation of indirect restorations was done following the manufacturer recommendations for IPS e.max CAD. The intaglio surface of each restoration was etched with buffered hydrofluoric acid gel 9.5% (Bisco porcelain etchant, Bisco, USA) for 20 s followed by rinsing for 20 s and another 20 s of air drying. A thin layer of silane primer (Porcelain primer, Bisco, USA) was applied to the etched surface and left for 30 s then air dried. Prepared teeth were etched with for 30 s with 37% phosphoric acid gel (Etch-37, w/BAC, BISCO Inc, USA), rinsed and air died. Followed by application of bonding agent and light polymerization for 10 s. All restorations were adhesively cemented to their corresponding teeth samples using dual-cure luting resin cement (Duo-link Universal ™, Bisco Inc, Schaumburg, USA) and a custom-made loading device was used to apply a vertical load of 1 kg. Initial curing for 3 s was done followed by removal of excess cement and 40 s curing on each surface. Finally, all samples were incubated in distilled water for 24 h before thermal cycling.

Marginal gap assessment

Marginal gap assessment before thermal cycling was done with direct viewing method using stereomicroscope (Leica MZ6, Leica Microsystems, ETH Zurich, Switzerland) with 30x magnification [18,19,20,21]. High definition digital camera was utilized for capturing images (Leica MC190 HD, Leica Microsystems, Switzerland). Four images were obtained for each sample (one for each axial surface). Images were uploaded to the image analysis software (Image Pro-plus V.6) for marginal gap measurement. Vertical marginal gap along each restoration surface was measured at 3 equal distance landmarks, Fig. (2). Three measurements were taken at each location [20]. All data were collected and tabulated to be used later.

Thermal cycling

All teeth were subjected to 5000 thermal cycles in masticatory simulator (Robota automated thermal cycle; BILGE, Turkey), equivalent to 6 months. In order to mimic the temperature changes occurring in the oral cavity, dwell periods in each water bath were set at 25 s with 10 s lag time. The minimum temperature was 5 °C. The maximum temperature was 55 °C [20].

Marginal gap was assessed for all samples after thermal aging with the same method and at the same pre-aging sites, Fig. (3)

Fig. 2

Stereomicroscopic images showing marginal gap measurement in different groups before thermal cycling. A: Group R, B: Group O, and C: Group C

Fig. 3

Stereomicroscopic images showing marginal gap measurement in different groups after thermal cycling. A: Group R, B: Group O, and C: Group C

Fatigue resistance test

Resistance to fracture under cyclic loading for all three study groups was done with stepwise stress fatigue method, using a computer-controlled testing machine (Model 3345; Instron Industrial Products, Norwood, USA) with 5kN load cell. Each sample was secured through the acrylic block to the lower part of the testing machine. A metallic stylus with 5 mm diameter of a spherical head and attached to the movable upper part was used to apply compressive load in a vertical direction to the occlusal surface of each sample. After simultaneous contact of the buccal and palatal cusps of the tooth with the spherical head was verified, 0.2 mm thick tin foil was inserted between the stylus head and the tooth to decrease the stress concentration and to mimic the presence of food. The cyclic loads to a maximum of 5000 cycles were applied to each sample at a crosshead speed of 0.5 mm/ minute and 1.6 Hz frequency. Each sample was subjected to a prescribed number of cycles at each of a sequence of increasing stress levels, until failure occurred. At the start of stepwise stress test, a load level below the expected fatigue failure for the material was selected (40% of the average static fracture force). The maximum number of cycles at each step load was set in 1000 cycles. If the specimen survived 1000 cycles, the stress level was raised by a constant load increment (100 N) in the same specimen. 200 N were used as initial load, followed by successive steps of 100 N each [16]. The load and number of cycles required for failure in each sample were recorded using computer software (Bluehill Lite; Instron Instruments). Calculation of the maximum fatigue load (LE) supported by each specimen was done according to the equation represented by Nicholas T [22]:

Where: L0 = Previous maximum fatigue load which did not cause failure, ∆ L = Load step increase, Nfail = Number of cycles till failure (L0 + LE), Nlife = Defined cyclic fatigue life (5000 cycles).

The number of intact samples and samples that failed were counted with each load step to calculate the survival probability (%) for each group.

For failure mode analysis, representative samples were examined with stereomicroscope at 10x magnification. Failure modes were categorized into three types according to their type and location: Type I (repairable fracture involving restoration only), Type II (repairable fracture involving restoration and tooth above CEJ), and Type III (Catastrophic, non-repairable fracture involving tooth and restoration below CEJ) [14].

Statistical analysis

Statistical analysis was performed with SPSS 16 ® (Statistical Package for Scientific Studies), Graph pad prism & windows excel. Data exploration was done with Shapiro-Wilk test and Kolmogorov-Smirnov test for normality which revealed that all data originated from normal data. For marginal adaptation, comparison between the three study groups was performed by One Way ANOVA test followed by Tukey`s Post Hoc test for multiple comparison. Comparison between before and after thermal cycling was performed by using Paired t test. For fatigue resistance, comparison of the fatigue failure load between the study groups was done by One-Way ANOVA test, followed by Tukey`s Post Hoc test for pairwise comparison. Life table survival analysis was performed to assess the survival probability (%) of samples in each group at each time interval (represented by each load step). Failure mode analysis (%) comparison was done by Chi-square test. All results were presented as frequency and percentages and all comparisons were performed by using the Chi-square test. The significance level of p < 0.05 was set for all tests.