Influence of splinted vs. non-splinted scan bodies on digital impression accuracy in all-on-four implant rehabilitation: a comparison with conventional open-tray impressions

This in vitro study aimed to evaluate and compare the accuracy of conventional and digital implant impression techniques using different intraoral scanners, implant systems, and splinting protocols in full arch implant rehabilitation. Two mandibular epoxy-resin models were fabricated and restored with four implants placed according to the All-on-Four concept using two different implant systems (Straumann implants with PEEK scan bodies and Megagen implants with titanium scan bodies). Digital impressions were obtained using three intraoral scanners (TRIOS 5, CEREC Primescan, and Virtuo Vivo), with the scan bodies scanned under both splinted and non-splinted conditions. In addition, conventional impressions were made using the splinted open-tray technique. All impression datasets were converted into STL files and superimposed onto digitally acquired reference models obtained with a laboratory scanner using best-fit alignment in Geomagic Control X 2022 v1.0.70 software. Trueness was evaluated by comparing each scan or impression dataset with the reference dataset, whereas precision was assessed by comparing all possible pairwise datasets within each group. Root-mean-square (RMS) deviation values were calculated for both trueness and precision. Statistical analyses were performed using a statistical software package (IBM SPSS Statistics, v20.0; IBM Corp., Armonk, NY, USA) at a significance level of α = 0.05. RMS values were significantly influenced by implant system, scanner type, and splinting protocol ( p < 0.05). Straumann scan bodies fabricated from PEEK demonstrated significantly lower RMS values (78.0 ± 34.7 μm) compared with Megagen titanium scan bodies (97.0 ± 34.3 μm) ( p = 0.001). Non-splinted digital impressions exhibited significantly higher trueness (67.8 ± 24.0 μm) than splinted impressions (118.5 ± 24.7 μm) ( p < 0.001). Significant differences were also observed among the intraoral scanners ( p < 0.001), with TRIOS 5 showing the lowest RMS values (78.1 ± 29.7 μm), followed by CEREC Primescan (87.1 ± 30.5 μm), while Virtuo Vivo demonstrated the highest RMS deviations (114.3 ± 35.0 μm). The conventional splinted open-tray technique yielded the lowest RMS values overall, with the Straumann open-tray group presenting the highest trueness (45.6 ± 9.5 μm). Although the main effects of implant system, scanner type, and splinting protocol were statistically significant, their combined effects varied according to the evaluated conditions rather than demonstrating a uniform interaction across all groups. Precision analysis supported the trueness findings, showing significantly higher precision in the non-splinted groups than in the splinted groups ( p < 0.001). Implant brand and scanner/impression type also significantly affected precision values ( p = 0.008 and p < 0.001, respectively). Within the limitations of this in vitro study, the accuracy of digital implant impressions was influenced by scanner type, implant system, and splinting protocol. Non-splinted digital impressions generally demonstrated higher trueness and precision than splinted impressions, particularly when TRIOS or CEREC scanners were used in combination with PEEK scan bodies. Nevertheless, the conventional splinted open-tray technique yielded the lowest RMS deviations overall, supporting its continued reliability as a reference method for full-arch implant impressions. These findings suggest that digital impression accuracy depends on specific factor-related effects rather than a single uniform protocol.