Effects of different loading approaches for in vivo-like Finite-Element (FE) models of total knee replacements on stress distribution and failure probability of ceramic tibial components

Abstract-Text (inkl. Referenzen und Bildunterschriften)

Introduction

In vivo-like FE studies are used during the development of joint replacements to evaluate risks of in vivo failures. In vivo data often used to derive loading conditions were published for various activities¹. This data was derived for a medium implant size of patients with a mean BMI of 30. Thus, the effects of two loading approaches were compared in terms of 1^st principal stress distribution and prediction of failure probability for ceramic tibial components.

Methods

Quasi-static FE models (femoral, tibial component, insert, bony structures, bone cement) of four sizes, were loaded with Peak100 data as reference models¹. Here, only moments were component-size scaled and a constant force was applied over the sizes. In comparison, a component-size scaled moment and force approach was used with increasing loads and moments from small to large sizes. This approach considers highly obese patients (BMI40). Failure probability was calculated based on analytical fracture mechanical algorithm.

Results

Using the scaled loading approach, the tibial components were more homogenously stressed over the portfolio instead of showing excessive stresses for smaller sizes and low stresses for larger sizes. Hence, worst-case size selection in terms of the largest failure probability was affected by the different loading approaches. The reference model revealed the smallest component size as worst case while the component-size scaled approach detected a large-mid size.

Summary

For risk mitigation of in vivo failures of knee implants, scaled moments and scaled forces shall be considered to avoid unrealistic high loading of small components and insufficient loading of large components. Especially for ceramic knee implants, the scaled loading approach should be considered as the failure probability analysis may reveal a different output in terms of worst-case sizes.

References

¹ Bergmann et al. PLoS ONE 2014 9(1):e86035