Abstract
Summary Statement
A population based finite element study that accounts for subject-specific morphology, density and load variations, suggests that osteoporosis does not markedly lower the mechanical compliance of the proximal femur to routine loads.
Introduction
Osteoporosis (OP) is a bone disease defined by low bone density and micro-architectural deterioration. This deterioration is neither uniform nor symmetric at the proximal femur. Evidence from analyses performed at the tissue level suggests that the cortical shell at the femoral neck is thinner in OP patients, especially in the superior regions, but not in the infero-anterior ones [Poole, Rubinacci]. Analogously, OP femurs show a higher anisotropy of the trabecular bone than controls [Ciarelli], suggesting a preservation of load bearing capacity in the principal loading direction vs. the transverse one. There is general consensus that the regions subjected to higher loads during walking, which is the predominant motor activity in the elderly, are mostly preserved. All these findings suggest that the OP femur should exhibit an almost normal mechanical competence during daily activities. This would be in accordance with the very low incidence of spontaneous fractures [Parker] and with the moderate fracture predictivity of BMD. Although reasonable, this hypothesis has never been tested at the organ level. Aim of the present study was to verify it with a population-based finite element (FE) study.
Patients & Methods
Whole femur Computed Tomography (CT) scans of 200 patients (115 women) with normal femoral anatomy were retrieved from a repository of the Istituto Ortopedico Rizzoli. The database is representative of an adult Italian population (mean 57yrs, range 23–84), and spans a wide range of morphological and densitometric characteristics (CT-simulated T-score of femoral neck BMD ranging from 1 to −4.6). Personalised FE models of all femurs were built from CT data using a validated procedure [Schileo]. A personalised estimate of the variability of loads acting on the proximal femur during normal walking (NW) and stair climbing (SC) was obtained by querying an indexed and searchable database of joint and muscle loads obtained from musculoskeletal models of 90 subjects. 78 possible loading combinations for NW and 50 for SC were defined for each subject, taking into account individual characteristics (height, weight, femoral antetorsion, CCD angle and neck length).
Risk of fracture (RF) was defined for each subject as the maximum principal strain / limit strain (1.04% compressive, 0.73% tensile) ratio over the whole loading spectrum.
Results and Discussion
No fracture was predicted by the FE models throughout the entire population, yielding an average safety coefficient of between 4 and 5, which is consistent with experimentally determined failure loads in the single leg stance configuration (around 11 BW [Cristofolini]). While a general inverse association was observed with R2∼0.2, no clear correlation was present between the fracture risk and the T-score. The hypothesis that OP does not macroscopically influence the mechanical competence of the femur for daily activities was therefore corroborated, suggesting that the highest risk of fracture in OP patients might be related to a lower OP induced compliance to accidental loads.
