The study groups for this study are presented in Table 1. Two typodont mandibular casts (Pro 2001-UL-SPFEM-32 typodont model, Nissin Dental Products Inc., Kyoto, Japan) with missing right first premolars were used to receive tooth preparation on the canine and second premolar teeth. For one of the typodont casts, the abutment preparation was performed with 6-dgree taper, 4.5 mm wall height, and 1 mm circumferential chamfer finish line. For the second typodont model, the abutment teeth were prepared with the same characteristics except having feather-edge (vertical) margin configuration [37]. The abutment preparation was performed by a single operator using a paralleling device (Dentalfarm A3006 B, Turin, Italy) and a straight hand-piece (Traus AT-II, Saeshin Precision Co., Daegu, Korea). Both prepared typodont casts were scanned (Ceramill Map 400, Amann Girrbach GmbH, Koblach, Austria). Based on the scan data, both preparation segments were printed (Phrozen shuffle, phrozen Technology, Hsinchu, Taiwan) using a printer resin (FTD Dentifix-3D LR, 3D printing resin, Lumi industries, Montebelluna, Italy). Then, the printed patterns were invested and casted to obtain two metal casts.
Forty impressions were made (Presigum, President Dental GmbH, Germany) and poured in type IV stone (Elite Master, Zhermack SpA, Italy) according to manufacturer’s instructions. After setting, each stone working cast was evaluated for any defects. All working casts were scanned (Ceramill Map 400, Amann Girrbach GmbH, Koblach, Austria). Then, all fixed partial dentures were designed using the CAD software (Ceramill Mind, Amann Girrbach GmbH, Koblach, Austria). A cement layer of 40 µm starting 0.5 mm from the margin, and connectors with 9 mm2 dimensions were chosen [38].
Fabrication of the restoration: For heat pressed zirconia-reinforced lithium disilicate fixed partial dentures (HL and VL groups), a pattern was printed (Phrozen shuffle, phrozen Technology, Hsinchu, Taiwan) using a printer resin (FTD Dentifix-3D LR, 3D printing resin, Lumi industries, Montebelluna, Italy) for each fixed partial denture. The resin patterns were sprued and invested, and then the pressing procedures were performed following the instructions of the manufacture using a heat pressed zirconia-reinforced lithium disilicate ingots (Vita Ambria, Vita Zahnfabrik, Bad Säckingen, Germany). Finally, each fixed partial denture was finished and cleaned before being subjected to glaze firing following the instructions of the manufacture. For zirconia fixed partial dentures (HZ and VZ groups), the restorations were dry milled from a zirconia disc (Zolid fx, Amann Girrbach GmbH, Austria) using the CAD-CAM milling machine (Ceramill motion II, Amann Girrbach GmbH). The milled zirconia restorations were sintered (Ceramill Therm 3, Amann Girrbach, Austria), then glazed based on the recommendations of the manufacturer.
Micro-CT scanning: Each fixed partial denture was seated on its respective metal model with finger pressure. To prevent movement of the restoration, utility wax was placed on the facial and lingual margins of each restoration. To evaluate the fit, each fixed partial denture was scanned by micro-CT equipment (SkyScan 1173 micro-CT scanner, Bruker, Kontich, Belgium). The scanning parameters were: x-ray source voltage of 40-130kV, source current of 61 µA, x-ray detector distortion-free flat panel sensor of 2240 x 2240 pixels and scan duration of 21 minutes per specimen. Projected images were reconstructed using special software (NRecon, SkyScan, NV, Belgium). The cross-sectional micro-CT images and the region of interest were obtained using the machine software (SkyScan Data Viewer 1.5.0.0, SkyScan, NV, Belgium). To evaluate the fit of each fixed partial denture, ten measurement sits were used: marginal gap (perpendicular measurement from the internal surface of the retainer to the margin of the die), chamfer area (800 µm occlusal to the margin of the die), axial wall (internal fit at the midpoint of the axial wall), axio-occlusal transition area (transition from the occlusal surface to the axial wall), occlusal area (500 µm from the axio-occlusal angle). Additionally, the horizontal marginal discrepancy was analyzed from cross-sectional images taken at the horizontal plane (Fig. 1).
Statistical analysis: Analysis of data was carried out with SPSS version 26. Testing the normality of data was done with Kolmogorov-Smirnov test and Levene test. Continuous variables were described as mean, standard deviation, minimum, median and maximum. Student t-test was used to compare two groups when data were parametric, while non-parametric data were compared by Mann-Whitney U median for non-parametric data. Two-way analysis of variance (ANOVA) was used to examine the influence of two different categorical variables (preparation and material) on dependent variables.