Is remote active feedback gait retraining comparable to in-person retraining 2 years post anterior cruciate ligament reconstruction

Purpose: Anterior cruciate ligament (ACL) tear greatly increases the risk of osteoarthritis, a leading cause of chronic pain and disability. At 2 years post ACLR, altered knee mechanics have been associated with structural changes and longitudinal patient-reported outcomes (PROs). Specifically, increased knee adduction moment (KAM) correlated with cartilage deep tissue matrix degeneration, and greater KAM and flexion moment (KFM) predicted worse PROs in 8 years. In-person laboratory based active feedback gait retraining to produce medial weight shift under the foot has been shown to reduce KAM without increasing KFM in healthy and osteoarthritic knees. This intervention has not been tested in ACLR knees. Recent advancements in wearable technologies render active feedback gait retraining feasible in natural environments. It is unknown whether in-person and remote gait retraining would have similar effects on the KAM and KFM. Thus, this study tested the hypotheses that (1) changes of KAM and KFM after in-person and remote active feedback gait retraining would not differ and that (2) the KAM would reduce without altering the KFM before and after 8 weeks of active feedback gait retraining. Methods: Twelve subjects (2.1 ± 0.3 years post primary unilateral ACLR) participated in this IRB-approved study. Between Jul 2019-Feb 2020, 6 subjects conducted 8 weeks of in-person active feedback gait retraining for 30 minutes/week with research staff. These subjects received haptic feedback generated from a vibratory motor and based on pressure data from a single-unit force sensor placed under the lateral rearfoot in their shoes. Between Feb-Jul 2021, 6 subjects conducted 8 weeks of remote active feedback gait retraining for 10 minutes/day and 7 days/week in their homes. They received active feedback consisting of an audible tone generated by a wireless insole (Moticon) and a custom smartphone app. Active haptic or auditory feedback was generated whenever the lateral rearfoot pressure exceeded an individualized training threshold. After the baseline gait test, all subjects explored training thresholds incrementally to shift weight medially to stop the active feedback, and the training thresholds were finalized according to individual results. One training session each week was conducted with research staff in the lab for the in-person group or on Zoom where the staff could see at least 3-4 steps of frontal view gait for the remote group. During the weekly observed training sessions, all subjects received verbal instructions to correct observable speed differences and gait asymmetries. Subjects returned to the lab for another gait test upon completion of the 8-week training. For all gait testing, subjects performed three walking trials at self-selected normal, slow, and fast speeds. Kinematic and kinetic data were acquired using a standard 3D motion capture system (Qualisys & Bertec) with the point cluster technique. Knee moments were calculated in BioMove (Stanford) using an inverse dynamic approach, expressed as external moments relative to the tibial frame, and normalized to %BW×Ht. We compared the two gait retraining approaches based on the KAM and KFM changes using independent t-tests. The KAM and KFM before and after training were compared using paired t-tests. Results: Baseline demographics, gait, and PROs were not different, except that the remote group was younger than the in-person group. No differences were observed in changes to KAM and KFM between In-person and remote active feedback gait retraining groups (p > 0.05 for all speeds; Fig 1A), supporting that both in-person and remote active feedback gait retraining similarly alter the KAM and KFM. The two groups were thus combined for testing the second hypothesis. The KAM1 reduced after active feedback gait retraining (Fig 1B), with a significant difference at slow speed (-0.25 [-0.46, -0.04]%BW×Ht). Similar trends were observed at normal speed (-0.12 [-0.24, 0.01] %BW×Ht) and at fast speed (-0.33 [-0.66, 0.00]%BW×Ht). The KAAI significantly reduced at slow (-0.12 [-0.20, -0.05] %BW×Ht×sec, p = 0.004) and normal speed (-0.08 [-0.14, -0.03]%BW×Ht×sec, p = 0.006). Significant changes in KAM2 and KFM were not observed at all speeds. Conclusions: This initial study shows that remote active feedback gait retraining had similar influence on altering the KAM and KFM as the in-person approach. Furthermore, 8 weeks of active feedback gait training reduced the KAM without significantly changing the KFM in human subjects who are 2 years post ALCR. The combined use of wearable technology and telecommunication permitted researchers to deliver and monitor gait retraining outside the lab. Additional studies involving additional subjects for longer periods of time are needed to test if the removal of in-person training location and schedule improves patient compliance and retention of the gait modification.