This study was conducted to investigate the effect of cement spacer on fit accuracy and fracture strength for 3-unit and 4-unit zirconia frameworks. The null hypothesis was rejected as cement spacer had a significant effect on marginal and internal adaptation and fracture strength of zirconia frameworks. In this study, fit accuracy and fracture resistance were assessed, as both are critical for long term success of fixed restorations, and both can be affected by cement thickness [3,4,5,6,7]. The cement spacer settings assessed in this research were 80 μm, as recommended by the CAD software program manufacturer, 30 μm and 50 μm as recommended settings in the literature [22,23,24,25,26,27]. Marginal and internal adaptation were assessed in this study with 2 different methods, one is the validated replica method [29,30,31,32]. One of the disadvantages of the replica method is the limited measuring points for marginal and internal gaps on the cement layer represented by the silicone replica [35]. The other method for assessment in this research was a modification of the replica technique by using an image analysis software with the same replica images [34]. The advantage of this method is that unlike the conventional replica method the whole total cement area was measured. The results for total cement gap in µm (average of marginal, mid-axial, axio-occlusal, and occlusal gaps) measured with conventional replica method was consistent with total cement area in mm2 measured with the modified method in the current study. Consequently, the modified replica method used in this research can be considered as a reliable method for similar future research. Both 3-unit and 4-unit frameworks had significantly better adaptation with a 50 μm cement spacer, followed by 80 μm and 30 μm cement spacers. In addition, the specimens with 50 μm spacer had more even silicone layer than the other spacers. This finding might be attributed to sintering shrinkage and deformation of zirconia frameworks that could not be compensated with a 30 μm cement spacer and lead to improper seating and poor fit accuracy [3, 6]. On the other side, an 80 μm cement spacer might have created excess space for cement (silicone), increased hydrodynamic pressures, and improper seating and adaptation of the frameworks [36]. Previous studies reported a marginal gap mean value for 3-unit zirconia FPDs ranging between 20 ± 5 μm and 106 ± 45 μm, a mean internal gap value between 30 ± 13 μm and 134 ± 47 μm [9, 12,13,14]. For 4-unit zirconia FPDs, Similar studies reported a marginal gap mean value for 4-unit zirconia FPDs ranging between 63 ± 36 μm and 141 ± 193 μm, a mean internal gap value between 58 ± 35 μm and 165 ± 137 μm [6, 11, 18]. The results for marginal and internal fit in the current research agree with Grajower and Lewinstein [37], who recommended the 50 μm spacer for better fit accuracy of fixed restorations. In addition, the results of the current research are consistent withKale et al. [22], and Schriwer et al. [23], who reported that the 30 μm spacer jeopardized the fit accuracy of CAD-CAM zirconia.On the other hand, the results of this study disagree with Suzuki et al. [27] who reported better adaptation with 30 μm spacer compared to 45 μm. However, the spacer settings assessed in this study were different.

In the current research, fracture resistance was assessed by measuring failure load and running the Weibull statistics to obtain more reliable results [7, 16]. Before failure load testing, the specimens were cemented with glass ionomer cement as performed in previous similar studies to simulate clinical routine [7, 7,8,9,10,11,12,13,14,15,16,17, 19, 20]. For fracture strength, both 50 μm and 80 μm had similar results, and both had significantly better fracture strength compared to 30 μm spacer for 3-unit and 4-unit specimens. This might be attributed to the ununiform cement layer with the 30 μm spacer that increased sintering stresses concentration within the specimens [4, 5]. The results for failure load in this study were within the range of the values reported in the literature [7,8,9,10,11,12,13,14,15,16,17, 19, 20]. The findings of this study are consistent with Rezende et al. [5], who reported that thicker cement space associated with misfit increased stress concentrations for zirconia crowns. On the other hand, the results of this study disagree with Schriwer et al. [23], who reported no significant effect for cement space on fracture strength of zirconia crowns. Unfortunately, the authors are unaware of available studies on the effect of cement spacer on zirconia FPDs to compare with the obtained results in the current research.

In Weibull statistic, shape parameter (m) indicates the predictability of fracture as a result of material flaws or defects, while scale parameter (σ0) indicates the characteristic fracture strength for 63.21% of the specimens [38]. No significant difference was found in the Weibull parameters between different spacers for 3-unit and 4-unit specimens. The Weibull distributions displayed a considerable overlap between 50 μm and 80 μm specimens, which was consistent with the results for fracture strength. The 3-unit specimens had significantly higher Weibull parameters than 4-unit specimens regardless of spacer setting, which agreed with the fracture strength results.

The results obtained in this research for Weibull parameters were consistent with the values reported in the literature [7, 16]. Similarly, the mode of failure in this study is consistent with the literature where the fracture initiated at the connectors or at the point of load application [4, 7, 15,16,17,18,19,20]. The mode of failure of specimens reflects the results of internal adaptation where the 30 μm and 80 μm spacers produced thick and uneven cement layer with more stress distribution and crack propagation within the axial walls.

For both adaptation and fracture resistance, 3-unit frameworks produced significantly better results compared to 4-unit frameworks regardless of spacer settings. As a result of increased volume of zirconia for 4-unit specimens, the sintering shrinkage and deformation might have increased with decreased adaptation and fracture resistance compared to 3-unit specimens [11].

The limitations of this study include the steel master dies used, which did not simulate dentin resiliency. However, this set up was selected for the purpose of standardization which would be difficult with natural teeth as abutments. In addition, the small sample size, as the fracture of dental ceramics is probabilistic, a larger sample size can be useful to reach more reliable results for future investigations. In addition, future research on monolithic zirconia FPDs with different cement spacer settings is recommended.