Moreover, considering the possibility of error while using an implant placement guide, we created many FEA models for each placement to compare the accuracy between the same placement models used with different FEA models.

Most studies verifying the usefulness of offset placement used a single technique for analysis [8–19]. Therefore, the results regarding the usefulness of offset placement vary depending on the techniques used in the studies, and hence, the issue remains controversial. The present study aimed to evaluate the usefulness of offset placement objectively by using two analyses—experimental and FEA models—and to verify the validity of each analysis. The limitation with experimental models is that it is not possible to observe internal stress, while FEA models have the disadvantage that it is required to confirm the agreement of the results with those obtained from the oral cavity of the living body. In the present study, we believe that it was possible to evaluate the usefulness of offset placement objectively by using multiple analyses so that the disadvantage of each analysis was compensated for by the other.

There is reportedly a concentration of stress in the surrounding cortical bone when occlusal load is applied to an implant [22, 25, 26]. Based on this, the strain gauges were applied to the peri-implant bone surface in the present study as well.

The experimental models and FEA models were similar and exhibited the same trends for compressed displacement based on differences in loading sites. In all placements, central loading resulted in the least compressed displacement and buccal loading the greatest. Thus, compressed displacement exhibited the same trend in the experimental models and FEA models. The results of both models may be reliable. When the effects of offset placement are considered, there is the concept of the load-supporting area (Fig. 15) put forth by Sato [27]. The load-supporting area is the area surrounded by the lines connecting the implant peripheries, and if a loading point falls within this area, there is little lateral force on the implant body. In the L-offset with buccal loading and the B-offset with lingual loading conditions, where the loading points are close to the load-supporting area, there is significantly less force than there is with straight placement (P < 0.05). Compared to the experimental models, the FEA models showed 10 μm less compressed displacement, but this may be the influence of the minute perturbations caused by differences in the constraint conditions in the artificial mandible bottom. The ability to completely reproduce the behaviors of experimental models with the FEA models would produce results with better approximation.