Towards understanding the ankle torque conundrum in isokinetic dynamometry: mechanical validation of dynamometer...

Isokinetic dynamometers (IKD) are widely used to measure passive and active torque generated at the joints, but active ankle plantarflexor torques measured with IKD are substantially lower than those derived from inverse dynamics (ID) for reasons that remain unclear. We conducted mechanical testing to investigate whether this discrepancy reflects intrinsic mechanical error in IKD measurement. A linear actuator-force plate setup was used to calculate torques via ID and compare them with IKD (Con-Trex) torques measured with ankle pedal attachment under four conditions that varied the vertical and lateral distances between the point of force application on the pedal and the IKD axis. Bland-Altman analysis showed small systematic bias across testing conditions (-2.97 to 2.30 Nm), but a significant proportional bias in all configurations (p < 0.001): IKD underestimated torque by ∼ 3 Nm at levels relevant to maximal plantarflexor effort and overestimated torque by 1-3 Nm at levels relevant to passive torque measurement. Reducing the lateral distance of force application by 4.5 cm reduced the external force required to produce the same IKD torque by 9% and substantially reduced off-axis torque. These findings indicate that mechanical bias alone is unlikely to explain the large IKD-ID discrepancy previously reported for human maximal plantarflexor torque, although it may be relevant for passive torque measurement. The greater off-axis loading observed with a larger lateral force-application distance may affect torque transmission within the dynamometer. Whether off-axis loading, mechanical compliance, or other factors contribute to the IKD-ID discrepancy during human testing remains to be determined.