Can individuals with knee osteoarthritis learn a gait modification after 3 weeks of unsupervised gait retraining

Abstract

Purpose: The knee adduction moment (KAM) is a surrogate measure of mediolateral loads across the tibiofemoral joint and is associated with radiographic severity and progression of knee osteoarthritis (OA). Recent literature suggests that healthy individuals can reduce KAM using auditory feedback from a lab-based pressure detecting shoe insole to reduce KAM. This clinical trial (NCT02955225) employs a novel, clinically feasible gait retraining protocol for patients with medial knee OA using pressure-based auditory feedback (PBF) generated from a wireless shoe insole (OpenGo, Moticon) and a smartphone. The training goal for following the auditory cues is to medialize foot center-of-pressure. We recently reported that patients with knee OA can respond to such clinically appropriate insole-driven training, subtly altering their walk to reduce KAM after a single baseline training visit. Here we explore how the same patients with medial knee OA kinetically responded after 3 weeks of unsupervised insole-based feedback training. Methods: In this IRB approved study, individuals with mild to moderate clinical and radiographic medial knee OA participated in two study visits over 3 weeks. During the baseline visit, subjects first walked in a standardized, flexible shoe (FlexOA, Dr. Comfort) at their comfortable pace (Week 0-Normal). Next, they trained with PBF from the wireless shoe insole and smartphone. To medialize foot center-of-pressure, they practiced to limit lateral plantar pressure to 85-95% (training parameters that were individualized based on each patient’s capability/tolerance at baseline). After about 30 minutes of in-lab training, subjects walked with the gait modification while receiving auditory feedback (Week 0-Feedback). Only walking trials in which participant successfully responded to the feedback were recorded. Subjects subsequently trained at home with insole-generated feedback for 3 weeks (training program: 5 min, 3x daily, 6 days per week), after which they returned for their Week 3 visit. During this visit, patients first walked with their normal, non-cued gait (Week 3-‘Normal’) and then with the gait modification they had been practicing over the last 3 weeks (Week 3-VerbalCue). It is important to note that no auditory feedback from the smartphone was provided during the Week 3 visit. For each visit, paired t-tests compared KAM, other knee moments, and spatiotemporal variables of the normal and modified conditions. Results: Data from 17 subjects were available and analyzed (14F/3M, 63 ± 10 years, 9 KL-2/8 KL-3). BMI remained unchanged between visits (30.4±5.9 kg/m2 & 30.3±5.9 kg/m2, p = 0.993). Peak knee moments and spatiotemporal variables are summarized in Table 1. At the baseline visit, subjects reduced peak KAM of midstance (KAM1) (p = 0.019, Fig 1) when feedback was provided, along with a reduction in speed (p = 0.035) and cadence (p = 0.032). At the Week 3 visit, after having trained with PBF for 3 weeks, KAM values were unchanged compared to baseline (W3-‘Normal’ vs. W0-Normal, p = 0.331). However, after being instructed to reproduce the gait modification, subjects walked with a significantly lower KAM1 (p = 0.026). This KAM reduction was attained without any significant difference in spatiotemporal variables, but was similar to the reduction obtained with auditory feedback during the baseline visit. Conclusions: After 3 weeks of training, individuals with knee OA had learned a gait modification that reduced peak KAM. No auditory cues were necessary to help in the execution of the modification. This is exciting as it suggests that patients can reproduce learned gait modifications without being reliant on feedback cues. However, their normal gait at the Week-3 time point did not differ from baseline. This suggests that patients achieved “motor performance” but not “motor learning” after 3 weeks of PBF. It is likely that additional training is required, including personal therapy sessions educating patients on the relevance of their achievements. This may allow transfer of their newly learned gait pattern into a normal habit. http://www.abstractsonline.com/admin/Report/abstractProofReport.aspTabled 1Table 1 Peak knee moments and spatiotemporal varialbesW0-BaselineW0-FeedbackPaired t-testW3-'Normal'W3-VerbalCuePaired t-testMean (SD)Mean (SD)p-valueMean (SD)Mean (SD)p-valuePeak knee moments (%BW*HT)Flexion1.845 (1.022)1.701 (1.091)0.5321.792 (1.089)1.588 (1.075)0.083Extension2.660 (0.905)2.510 (1.077)0.1942.869 (0.892)2.680 (0.998)0.195KAM12.877 (0.756)2.665 (0.837)0.0192.955 (0.852)2.774 (0.894)0.026KAM22.191 (0.946)2.245 (0.999)0.3602.321 (0.936)2.313 (0.945)0.898External Rotation0.146 (0.094)0.136 (0.089)0.3960.136 (0.086)0.139 (0.088)0.805Internal Rotation0.953 (0.311)0.946 (0.334)0.8131.025 (0.354)0.994 (0.404)0.311Spatiotemporal VariablesSpeed (m/s)1.210 (0.203)1.143 (0.206)0.0351.235 (0.203)1.174 (0.201)0.116Stride length (m)1.251 (0.175)1.243 (0.188)0.7831.291 (0.154)1.254 (0.170)0.169Cadence (strides/min)56.49 (5.14)54.58 (4.96)0.03257.26 (4.96)55.95 (4.70)0.087 Open table in a new tab