Experimental evidence of strain uniformity and organ-level attenuation of viscoelasticity in human femurs under physiological multiaxial loading.

Abstract

We experimentally evaluated whether physiological multiaxial loading produces comparatively constrained strain distributions in the human femur and examined how classical tissue-level strain-rate dependence translates to apparent organ-level behaviour. Eight cadaveric femurs from female donors aged 56-91 years were instrumented with ten rosette strain gauges. Hip contact force profiles for walking, stair ascent and descent, chair rise and sit, and stumbling were obtained from Orthoload, scaled to each donor's body weight, and applied using a six-degree-of-freedom hexapod robot with a hybrid FPGA-CPU load-control architecture. Low-speed force-control trajectories were replayed under position control at increasing speeds, with five cycles per speed. Low-speed load replication error and cycle-to-cycle repeatability were <3 N and <1%, respectively. Peak equivalent cortical strain ranged from 61 to 2201 με. Across activities, pooled median strain remained comparatively constrained, ranging from 508 με for chair rise/sit to 661 με for walking, with no significant activity effect and negligible effect size. Using the actual activity-specific frequency range, 0.015-0.58 Hz, apparent femoral stiffness increased logarithmically with loading frequency, corresponding to an equivalent exponent of 0.012. A sensitivity analysis based on the force-to-cortical-strain response yielded a smaller strain-normalised exponent of 0.004. Both values were markedly lower than the classical tissue-level exponent of approximately 0.06. These findings support cumulative strain uniformity principles in previous bone-adaptation theories while showing that tissue-level loading-rate dependence is attenuated at the whole-femur scale. The data also provides a benchmark for model validation and standardisation.