Prediction of proximal femur strength using a CT-based nonlinear finite element method: Differences in predicted fracture load and site with changing load and boundary conditions

The annual occurrence of hip fracture due to osteoporosis as of 2002 had reached 120,000 in Japan. The increase has been very rapid. From a biomechanical perspective, hip fractures are thought to be caused in real settings by different directions of loading. Thus, clarification of the loading directions under which the proximal femur is most vulnerable to fracture would be helpful for elucidating fracture mechanics and establishing preventive interventions. The purpose of the current study was to clarify the influence of loading direction on strength and fracture site of the proximal femur using the CT-based nonlinear FE method to determine loading directions under which the proximal femur is most vulnerable to fracture. Contralateral femora were analyzed in 42 women with hip fracture (mean age, 82.4 years), comprising 20 neck fractures and 22 trochanteric fractures. Within I week after fracture, quantitative CT of the contralateral femur was performed in each patient and 3-dimensional FE models were created. One stance loading configuration (SC) and four different fall loading configurations (FC) were assigned. Nonlinear FE analysis was performed. Differences in fracture loads depending on differences in loading direction were analyzed and correlations among fracture loads in different loading directions were assessed. Next, fracture sites were also analyzed. Mean predicted fracture load in the SC was 3150 N. Mean fracture loads were 2270 N in FC1, 1060 N in FC2, 980 N in FC3, and 710 N in FC4. The correlation between predicted fracture loads in SC and those in each FC was significant with a correlation coefficient of 0.467-0.631. Predicted fracture sites in the SC appeared at the subcapital region in all patients and were categorized as neck fracture. However, trochanteric fractures occurred in all fall configurations except FC1. In FC1, a significant correlation was seen between real fracture type and predicted type. The current investigation could contribute to the acquisition of useful knowledge allowing the establishment of more efficacious means of preventing hip fractures. (C) 2009 Elsevier Inc. All rights reserved.