CAN STEREO-LITHOGRAPHICALLY PRODUCED ARTIFICIAL BONES BE USED FOR PRECLINICAL MECHANICAL TESTING OF IMPLANTS?

Abstract

Introduction: Bio mechanical tests under realistic loading conditions of prostheses in bone can help to improve the design of joint implants. Cadaveric bones are most realistic but highly variable and difficult to obtain and conventional bone models have been used so far. Stereo lithography (SLA) techniques are used in industry to generate 3-D rapid prototypes. These techniques could serve to produce bones with complex geometries, but the material used is less stiff than cortical bone.

Aim: The purpose of the study was to answer the following two questions? 1. Does stability of and cortical strains around implants in SLA-made bones matched those of conventional artificial bones? 2. Whether increasing cortical wall thickness brings these variables closer?

Methods: Four artificial cortical shells of proximal tibiae were made from resin (SL5170, 3D systems Europe Ltd., Hemel Hempstead, UK) using SLA process. Two third generation large composite tibiae #3302 (Sawbones Europe AB, Malmö, Sweden) were chosen and the polyurethane foam that represents the cancellous bone was removed. All six cortices were filled with polyurethane foam (Tripor 224, ABL (STEVENS), Cheshire, UK) with an average compressive modulus of 53.9±7.2 SD MPa. The tibiae were prepared to receive a standard size cemented tibial tray for all models. The models were loaded with 100 cycles of 2000 N at 1 Hz along the longitudinal axis, separately on the lateral and on the medial condyle. Medial cortical strain and tray migration during load was determined.

Results: Cyclic loading gave a general pattern of cyclic movements, superimposed on a very small permanent movement. The first cycle gave most permanent displacement, after which further migration occurred at a decreasing rate. Permanent and cyclic migration of all four trays implanted in SLA-made tibiae fell within the range of those implanted in conventionally available tibiae. Strains at the proximal medial cortex were low and on the same order for all six tibiae. Strains more distally were approximately inversely proportional to the material stiffness and cortical thickness of the tibiae.

Conclusion: The study concludes that migration of tibial trays in all SLA models was with in the range of those in conventional models. Hence these models can be used to test early mechanical stability of joint implants despite their lower stiffness. The small difference may be related to load bearing mechanism of tibial trays which is largely through cancellous bone and not cortical bone. The low strains at the proximal cortex in this study also suggest that the cortex carried little direct load. The polyurethane foam representing cancellous bone in our study was identical for each tibia, which may explain that movements of the trays were comparable. Distal cortical strains reflected the stiffness of the tibiae and were directly influenced by cortical thickness.