Gauging force by tapping tendons? Inaccurately in the human patellar tendon

Accurately estimating in vivo tendon load non-invasively remains a major challenge in biomechanics, which might be overcome by shear-wave tensiometry. Shear-wave tensiometry measures the speed of mechanically induced tendon shear waves by skin-mounted accelerometers. To gauge the feasibility and accuracy of this novel technique, we obtained patellar tendon shear wave speeds via shear-wave tensiometry during sustained or ramp voluntary contractions of the knee extensors in two experiments (n = 8 in both). In experiment one, participants produced a constant knee extension torque of ∼ 50 Nm at five different knee joint angles (i.e. variable tendon load), which resulted in estimated patellar tendon forces between 1005 ± 6N and 1182 ± 16 N. However, wave speed squared did not correlate with estimated tendon force within participants (r_rm(31) = -0.19, p = 0.278). In experiment two, averaged correlation coefficients between normalized wave speed squared and torque of maximal and submaximal voluntary contractions across participants ranged between r = 0.43 and r = 0.94, while the time-varying correlation between these normalized signals ranged from r = -0.99 to r = 1.00. Further, the mean absolute errors (MAEs) between normalized wave speed squared and normalized torque across participants ranged between 0.03 and 0.54, which were larger than the MAEs between normalized torque and normalized summed EMG amplitude from the superficial quadriceps muscles (0.03-0.54 versus 0.06-0.26, respectively). In conclusion, there was no simple relation between shear wave speed squared and patellar tendon load, which severely limits the feasibility of shear-wave tensiometry for accurately estimating in vivo tendon load at the knee joint.