Tissue viscoelasticity is related to tissue composition but may not fully predict the apparent-level viscoelasticity in human trabecular bone – an experimental and finite element study

Trabecular bone is viscoelastic under dynamic loading. However, it is unclear how tissue viscoelasticity controls viscoelasticity at the apparent-level. In this study, viscoelasticity of cylindrical human trabecular bone samples (n = 11, male, age 18–78 years) from 11 proximal femurs were characterized using dynamic and stress-relaxation testing at the apparent-level and with creep nanoindentation at the tissue-level. In addition, bone tissue elasticity was determined using scanning acoustic microscope (SAM). Tissue composition and collagen crosslinks were assessed using Raman micro-spectroscopy and high performance liquid chromatography (HPLC), respectively. Values of material parameters were obtained from finite element (FE) models by optimizing tissue-level creep and apparent-level stress-relaxation to experimental nanoindentation and unconfined compression testing values, respectively, utilizing the second order Prony series to depict viscoelasticity. FE simulations showed that tissue-level equilibrium elastic modulus (E) increased with increasing crystallinity (r = 0.730, p =.011) while at the apparent-level it increased with increasing hydroxylysyl pyridinoline content (r = 0.718, p =.019). In addition, the normalized shear modulus g (r = −0.780, p =.005) decreased with increasing collagen ratio (amide III/CH) at the tissue-level, but increased (r = 0.696, p =.025) with increasing collagen ratio at the apparent-level. No significant relations were found between the measured or simulated viscoelastic parameters at the tissue- and apparent-levels nor were the parameters related to tissue elasticity determined with SAM. However, only E, g and relaxation time τ from simulated viscoelastic values were statistically different between tissue- and apparent-levels (