When a three-dimensional FEA is used to analyze the mechanics of peri-implant bone, it is ideal to construct an FEA model that approximates the material properties and structures of an actual mandible. Moreover, the results should be compared with the behavior of an implant in an actual mandible. However, in an actual oral cavity, individual differences exist resulting from bone morphology and physical properties; therefore, it is difficult to conduct experiments under constant conditions and to obtain results that can be applied to all individuals. In other words, to systematically analyze the mechanics of peri-implant bone, an artificial bone model in which individual differences can be eliminated is regarded as valid. The artificial mandibular bone used in this study was regarded as type II in the Lekholm and Zarb classification [18] and had been fabricated on the assumptions of having adequate bone quality, internal structure, and morphology for clinically valid implant therapy. It is difficult to say whether the experimental model was an ideal model because the experimental model has different material properties from those of an actual mandible. However, the purpose of this study was not to compare it with the behavior of an implant in an actual oral cavity; the purpose was to perform a comparison between an experimental model and FEA model to verify the reproducibility and validity of a three-dimensional finite element model. Therefore, we used artificial mandibular bone in the experimental model rather than an actual mandible.

In many reports on the three-dimensional FEA of implants, loading was carried out using a simplified FEA model in which the cancellous bone interior was regarded as a homogeneous body [19-23]. This is partly because X-ray CT imaging does not provide adequate resolution, and it is difficult for CT to accurately reflect the state of contact between the trabecular structure and the implant [24]. Therefore, it is inevitable that loading will be performed using a simplified FEA model in which the cancellous bone interior is regarded as being a homogeneous body, as was the case in the present experiment. FEA in industry has been utilized as a ‘rough analysis’ (first-order analysis) tool by simplifying the details in order to ascertain an overall tendency in the first stages of structural design [25]. Therefore, in the present study, a ‘first-order analysis’ was utilized to ascertain the behavioral tendencies of the implants as a first step before proceeding with an analysis of implant mechanics using a three-dimensional FEA. As has been performed in many reports on the FEA of implants, we verified the validity of the FEA models by studying the extent to which the actual behaviors were reproduced when the trabecular structure in the FEA models was simplified and compared with the experimental model.