Elsevier

Bone

Volume 64, July 2014, Pages 108-114
Bone

Original Full Length Article
Analysis of strength and failure pattern of human proximal femur using quantitative computed tomography (QCT)-based finite element method

https://doi.org/10.1016/j.bone.2014.04.007Get rights and content

Highlights

  • A reliable and efficient scheme was developed for failure analysis of human proximal femur using QCT-based finite element method.

  • Failure pattern and strength of femoral specimens were predicted using a strain energy density-based risk factor through linear analyses.

  • Ten femoral specimens were modeled, analyzed, and mechanically tested under eight different loading orientations.

  • Very good agreements were found between predicted and experimental results of specimens with marked differences in size and strength.

  • The average computational run-time was 5 min for an average element number of 197,000.

Abstract

This paper presents a novel method for fast and reliable prediction of the failure strength of human proximal femur, using the quantitative computed tomography (QCT)-based linear finite element analysis (FEA). Ten fresh frozen human femora (age: 34 ± 16) were QCT-scanned and the pertinent 3D voxel-based finite element models were constructed. A specially-designed holding frame was used to define and maintain a unique geometrical reference system for both FEA and in-vitro mechanical testing. The analyses and tests were carried out at 8 different loading orientations. A new scheme was developed for assortment of the element risk factor (defined as the ratio of the strain energy density to the yield strain energy for each element) and implemented for the prediction of the failure strength. The predicted and observed failure patterns were in correspondence, and the FEA predictions of the failure loads were in very good agreement with the experimental results (R2 = 0.86, slope = 0.96, p < 0.01). The average computational time was 5 min (on a regular desktop personal computer) for an average element number of 197,000. Noting that the run-time for a similar nonlinear model is about 8 h, it was concluded that the proposed linear scheme is overwhelmingly efficient in terms of computational costs. Thus, it can efficiently be used to predict the femoral failure strength with the same accuracy of similar nonlinear models.

Introduction

Hip fractures cause significant numbers of disability and mortality worldwide. The incidence of these fractures varies globally, e.g., the city of Shiraz in Iran is reported to have the highest incidence of hip fractures in Asia [1]. Previous studies estimated that the worldwide number of hip fractures was 1.3 million in 1990 and predicted an increase between 7.3 and 21.3 million by 2050 [1], [2], [3]. Thus, many researchers have tried to devise new methods for noninvasive assessment of the hip fracture risk, among which the QCT-based finite element analysis (FEA) has been recognized as a very reliable tool [4], [5], [6], [7], [8], [9], [10], [11], [12]. However, there are some limitations that should be resolved before this method can be clinically feasible. One major problem is the high computational time and cost for implementation of nonlinear FEA which has been shown to have the required accuracy [4], [6], [7], [8], [9], [13]. A typical QCT-based nonlinear FEA of human proximal femur requires at least 8–10 h of runtime, even by implementing advanced computational techniques [4], [7], [9], [13]. On the other hand, the linear FE analyses are very much faster, quite easier to apply, and can provide comparable predictions of the failure location and pattern for the proximal femur and vertebra [13], [14]. However, they require further data processing schemes for predictions of the failure load that may give less accurate results in certain cases. For instance, Keyak et al. proposed a linear FEA method for the prediction of the failure load in the sideways fall and stance configurations and argued that failure of a certain number of contiguous non-surface elements can represent the failure of the whole proximal femur [15]. On the other hand, Nishiyama et al. defined a constant yield strain level and suggested that the failure of a constant percentage of elements would represent the failure of proximal femur in the sideways fall configuration [16]. Nevertheless, we believe that the percentage or the number of critical elements should naturally be different for various specimens and loading orientations. Moreover, our previous studies on human vertebrae and proximal femur have shown favorable results in the prediction of the failure patterns using a strain energy density measure within the framework of linear QCT-based FEA [13], [14]. Thus, this study was aimed at development and experimental verification of a new method for fast and accurate prediction of the failure strength of human proximal femur through specific assortments and usage of high strain energy elements of the linear QCT-based FEA results.

Section snippets

Sample preparation

Ten fresh-frozen human femora, from 9 cadavers (5 male, 4 female, average age: 34 ± 16), were used in this study (see Table 1) (specimens 9 and 10 were a contralateral pair from a single cadaver).

The initial treatment of the specimens included standard excision, freezing, and bagging procedures by the Iranian Tissue Bank (ITB). Moreover, the death causes announced by ITB were the three categories of coronary failure, cerebral death, and fatal trauma. Nevertheless, the plain radiographs of the

Failure patterns

Fig. 3 shows the predicted and experimental failure patterns for three specimens (numbers 4, 5, and 9) under different loading orientations (see Table 1). It should be noted that in this figure only the screened critical elements are shown for clarity. For the specimen 4, the initiation of a trochanteric local failure was distinguished. The occurrence of a similar failure pattern at the same location (in the form of a distinct crack) was clearly visible on the experimental specimen (see Fig. 3).

Discussion

Many investigations have been carried out on non-invasive assessment of the hip fracture risk using the QCT-based finite element analysis (FEA). The valuable advantages of this method are the ability of development of very accurate image-based 3D models of the femoral geometry, plus a pointwise description of BMD-based material properties. In general, the results have shown that nonlinear FEA can predict the femoral strength with very good accuracy. However, several researchers have faced

Acknowledgments

The authors wish to thank Dr. Mohsen Sadeghi for his continuous help during several stages of this work. The valuable help of Dr. Mahdavi and Dr. Khodadadi from the Iranian Tissue Bank (ITB, Tehran University of Medical Sciences) in providing Cadaveric samples is highly appreciated. The QCT scans were carried out using the facilities of the Noor Clinic with the help and support of Dr. Akhlaghpour. The valuable help of Mr. Kargar (Materials Laboratory of Tarbiat Modares University) with the

References (24)

Cited by (33)

  • Investigation on distal femoral strength and reconstruction failure following curettage and cementation: In-vitro tests with finite element analyses

    2019, Computers in Biology and Medicine
    Citation Excerpt :

    Damage modeling was considered, using a traction-separation law, to investigate the debonding and damage initiation as well as damage propagation at the interface [45,46]. Nodes of the medial condyle, which were in contact with the 20 mm diameter actuator in the mechanical tests, were identified and an elastic modulus of 20 GPa and yield strength of 200 MPa were assigned to the elements containing those nodes to prevent severe distortion while being loaded as suggested in previous studies [33,36]. A displacement equal to 1 cm was applied to those nodes in the form of compression in the direction of the femoral shaft.

  • Linear and nonlinear analyses of femoral fractures: Computational/experimental study

    2018, Journal of Biomechanics
    Citation Excerpt :

    A group of 15 fresh-frozen human femora were used in this study (see Table 1). The group included 8 specimens from previous studies (Mirzaei et al., 2014), plus 7 new specimens. The initial treatment of the specimens included standard excision, freezing, and bagging procedure by the Iranian Tissue Bank (ITB).

  • Study of sex differences in the association between hip fracture risk and body parameters by DXA-based biomechanical modeling

    2016, Bone
    Citation Excerpt :

    After adjustment for BMD, impact force is positively correlated with FRI in both men (r = 0.83 , p < 0.001) and women (r = 0.81 , p < 0.001). A number of experimental and analytical studies have been performed to determine the range of impact force that can cause hip fracture [31,83–88]. However, the results of our study showed that the correlation between impact force and fracture risk was not strong before adjustment for BMD.

View all citing articles on Scopus
View full text