The failure of fixed prostheses has always been a concern for prosthodontists. One of the keys to the success of fixed restorations is the accuracy of impressions. Affected by complicated oral environment, the accuracy of both traditional and digital impressions carried out in clinic are generally inferior to those of in vitro trials or randomized controlled clinical trials. In this study, although there are more interference factors than other in vitro trials or clinical randomized controlled trials, the sample size is abundant here. Moreover, the sample in our study can reflect the clinical application effect completely, and the results can provide clear clinical reference value [29]. All study investigators, including prosthodontists, nurses, technicians, and quality inspectors, came from the same hospital and were trained and qualified to better reflect the actual application of the traditional and digital impressions. To our knowledge, few hospitals have quality inspectors who record and analyze the causes behind the failure of restorations.
Our results showed that the overall rework rate of digital impressions was 1.93%, which was lower than that of traditional impressions (2.31%). However, there was no significant difference. Previous in vitro studies have suggested that, direct intraoral scanning for partial-arch, or make impressions using silicone impression materials, or laboratory scanner scanning can meet the clinical requirements accurately and effectively [2, 16, 28]. Our results are consistent with those studies [19, 30]. The difference in failure rates between the two impression techniques is mainly due to the complex oral environment, including the presence of saliva, blood, tongue, and labial-buccal muscles, which could affect the accuracy of intraoral scanning and traditional impressions [28]. In addition, the cooperation of the patients and the technique of the operators may also have an impact. The adequacy of environmental light can particularly impact intraoral scanning.
As shown in Table 1, the rework rate of the traditional impression group did not increase with the number of abutment teeth of the restorations. On the contrary, the number of abutment teeth had a significant impact on the rework rate in the digital impression group. When the number of abutment teeth exceeded two (≥ 3 units), the rework rates of restorations doubled, and this situation was also reported in other studies. A study by Bi and his colleagues [11] suggested that there is no difference between the two impression techniques for short-span scanning. However, traditional impressions have been shown to have better accuracy than digital impressions for long-span scanning. Similarly, the review by Kihara and his colleagues [28] demonstrated a higher accuracy of the impressions with elastic impression materials, which did not fluctuate with the abutment teeth span.
We also analyzed the causes of rework (Table 2) and found that the rework rates of try-in failure due to abnormal occlusion (Group 1) existed no significant difference. If a patient had an abnormal occlusal position or excessively strong occlusal force during the intraoral scanning process, abnormal occlusion of the final restoration can occur [31]. Not only the inaccuracy during impression process, but also the high spots or deformation of the casts can possibly lead to abnormal occlusion of the traditional group [32]. In the present study, even though there was no significant difference, the rework rate in the digital impressions group was lower than that for the traditional group. In a study by Beck and his colleagues, a laboratory scan showed a more precise occlusal contact [33], which was in consistent with the results of our study.
During the digital impression process, incompatible edges can be caused by insufficient lighting, excessive blood or saliva, gingiva, and inappropriate drawing of edge lines of the softwares [28, 30]. Correspondingly, inaccuracies in impressions and casts could affect the marginal adaptation during the process of traditional impression, including visible bubbles and small defects [34, 35]. Different clinically acceptable marginal gap values have been reported, ranging from 39 to 150 µm [14, 15, 34, 36, 37]. Systematic reviews and meta-analyses of Tabesh [35] and Takeuchi [38] found that intraoral scanning showed better marginal adaptation than traditional impressions or desktop scanning for making fixed restorations. However, some investigators have found no difference in marginal adaptation among the several impression methods [39]. Generally, as same as the rework rate of try-in failure due to abnormal occlusion, the rework rate in the digital impressions group was lower than that for the traditional group, with no significant difference.
With regards to the reworked restorations due to unable positioning, the rework rate of try-in failure was significantly lower in the digital group than in the traditional group. The main reasons for failure due to unable positioning included deformed model, excessive undercut, and lack of a common insertion path. The models of oral scanning tended to deform with the expansion of the scanning range, whereas traditional casts did not [11, 16, 28]. Notably, the majority of cases in this study had a small scan range. Therefore, the difference in failure rate was unlikely to be influenced by the scanning range. In addition, if the technicians found significant abnormalities in abutment teeth during intraoral scanning, they asked the dentists to repeat the teeth preparation. It was more difficult to identify problems in time during the process of the traditional impression operation. As a result, there were fewer obvious defects in the preparations of the digital group, which may also be the main reason for its lower failure rate compared with the traditional group.
In summary, there were no significant differences in the rework rate between the digital and traditional impression groups, and both could meet the clinical requirements. However, in the present study, the sample size of the traditional group was much smaller than that of the digital group. Furthermore, there may be individual differences in dentists’ tolerance for failure rates, and more deeper analyses with additional clinical cases are needed.