Influence of the Angulation and Insertion Depth of Implants on the 3D Trueness of Conventional and Digital Impressions

Purpose: To study the influence of insertion depth and implant angulation on the 3D trueness of models obtained with different impression techniques.
Materials and Methods: Four different reference models (model 1: parallel, depth of 1.5 mm; model 2: parallel, depth of 4 mm; model 3: 20-degree angle, depth of 1.5 mm; and model 4: 20-degree angle, depth of 4 mm) of partially edentulous maxillae were generated by altering implant angulations and subgingival depths. All scans of reference models were done with a laboratory scanner, and obtained data were exported into standard tessellation language format to be used as virtual reference images. Impressions were obtained from each reference model via three conventional techniques (closed tray [CT], non-hexed open tray [NHOT], and hexed open tray [HOT]) and one digital technique (intraoral scanning [IOS]). A total of 160 impressions were made. The reference and experimental scan data were superimposed by using the best-fit alignment algorithm. Angular (AD), linear (LD), and 3D (RMS) deviations were computed, and obtained data were statistically analyzed.
Results: In premolar implant sites, AD and LD values were significantly affected by model type (P ≤ .001 for AD and LD) and impression technique (P = .001 for AD, P = .002 for LD). However, no significant interaction was detected (P = .703 for AD, P = .768 for LD). Model 1 (0.44 ± 0.25 for AD, 7.79 ± 6.29 for LD) and the NHOT technique (0.49 ± 0.43 for AD, 9.04 ± 8.14 for LD) exhibited the lowest mean AD and LD values. In molar implant sites, AD and LD values were significantly affected by model type (P ≤ .001 for AD and LD) and impression technique (P ≤ .001 for AD and LD), as well as by their interaction terms (P = .037 for AD, P = .005 for LD). Considering interaction terms, while the highest and lowest mean AD values were exhibited by IOS-model 4 (1.56 ± 0.25) and NHOTmodel 2 (0.46 ± 0.28), respectively, the highest and lowest mean LD values were exhibited by CT-model 4 (41.40 ± 14.48) and NHOT-model 2 (8.03 ± 4.86), respectively. RMS estimate values were significantly influenced by model type (P ≤ .001) and impression technique (P ≤ .001), as well as by their interaction terms (P = .019). The highest and lowest mean RMS values were exhibited by IOS-model 4 (70.02 ± 4.74) and NHOT-model 2 (25.96 ± 17.67), respectively.
Conclusion: In the case of angulated and deeply placed implants, splinted NHOT and HOT techniques can be recommended for better trueness.