Fabric–elasticity relationships of femoral head trabecular bone are similar in Type 2 diabetes and non-diabetic individuals

Type 2 diabetes (T2D) is a chronic disease leading to an elevated glucose level in the blood and increased fracture risk. Because T2D individuals tend to have normal to higher areal bone mineral density (aBMD) than healthy individuals, their fracture risk is often underestimated. As an alternative, high-resolution peripheral quantitative computed tomography (HR-pQCT) is an attractive tool to investigate bone morphology in vivo and estimate fracture risk. Based on HR-pQCT scans, bone strength can be estimated using micro finite element ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>μ</mml:mi></mml:math> FE) analysis or homogenized finite element (hFE) analysis. While <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>μ</mml:mi></mml:math> FE is computationally expensive, hFE provides an accurate estimation of bone mechanical properties within reasonable efforts. However, the hFE scheme is based on relationships between the local fabric (anisotropy) and elasticity. These relationships have been shown to hold for healthy controls as well as in the case of osteogenesis imperfecta. Nevertheless, whether these relationships are also valid for T2D-diagnosed patients remains unclear. Therefore, the present work aims to compare fabric-elasticity relationships between T2D and non-diabetic controls. The present study collected 56 trabecular bone cores from the femoral head of 28 T2D and 28 control donors. These cadaveric samples were scanned in a micro-CT system at an isotropic 14.8 <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>μ</mml:mi></mml:math> m voxel size. Three cubic regions of interest (ROIs) were selected in each scan. The resolution of these ROIs was coarsened by a factor of 4, mimicking clinical HR-pQCT resolution, and the ROIs were subsequently segmented. Standard morphometric parameters were computed from the segmented ROIs using medtool (v4.8; Dr. Pahr Ingenieurs e.U., Pfaffstätten, Austria). Additionally, their fabric tensor and their apparent stiffness tensors were computed. The ROIs were compared between T2D and control regarding their morphometric and mechanical properties. Finally, ROIs were matched between T2D and control for bone volume fraction ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>ρ</mml:mi></mml:math> ) and degree of anisotropy (DA). The matched dataset allowed the comparison of fabric-elasticity relationships between T2D and control samples. No significant difference was observed between T2D and control samples, both regarding their morphology and their mechanical properties. Specifically, fabric-elasticity relationships were shown to hold for both the control and the T2D groups. A comparison of the resulting exponents related to <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>ρ</mml:mi></mml:math> and DA has highlighted different trends but no important difference between T2D and control samples. In conclusion, trabecular bone architecture was similar between T2D and non-T2D donors. Additionally, fabric-elasticity relationships, i.e. morphology-mechanical relationships, are also similar between donors with and without diabetes. Accordingly, HR-pQCT-based hFE analysis could also be used for estimating the bone mechanical properties of T2D patients and for their fracture risk assessment.