Comparing walking with knee-ankle-foot orthoses and a knee-powered exoskeleton after spinal cord injury: a randomized, crossover clinical trial

Abstract

Recovering the ability to stand and walk independently can have numerous health benefits for people with spinal cord injury (SCI). Wearable exoskeletons are being considered as a promising alternative to conventional knee-ankle-foot orthoses (KAFOs) for gait training and assisting functional mobility. However, comparisons between these two types of devices in terms of gait biomechanics and energetics have been limited. Through a randomized, crossover clinical trial, this study compared the use of a knee-powered lower limb exoskeleton (the ABLE Exoskeleton) against passive orthoses, which are the current standard of care for verticalization and gait ambulation outside the clinical setting in people with SCI. Ten patients with SCI completed a 10-session gait training program with each device followed by user satisfaction questionnaires. Walking with the ABLE Exoskeleton improved gait kinematics compared to the KAFOs, providing a more physiological gait pattern with less compensatory movements (38% reduction of circumduction, 25% increase of step length, 29% improvement in weight shifting). However, participants did not exhibit significantly better results in walking performance for the standard clinical tests (Timed Up and Go, 10-m Walk Test, and 6-min Walk Test), nor significant reductions in energy consumption. These results suggest that providing powered assistance only on the knee joints is not enough to significantly reduce the energy consumption required by people with SCI to walk compared to passive orthoses. Active assistance on the hip or ankle joints seems necessary to achieve this outcome.

Figure 1

(a) Study protocol. (b) Representation of the data obtained from the GXT performed with an arm cycle ergometer. (c) Set up representation for the 6 MWT. (d) Marker protocol: Markers were placed at the trochanter, lateral side of the knee, lateral side of the ankle, hallux, and heel (achilles attachment). (e) Knee-ankle-foot orthoses with Walkabout. (f) Knee-powered lower limb exoskeleton; i.e., the ABLE Exoskeleton (ABLE Human Motion S.L., Barcelona, Spain).

Figure 2

(a) Distance covered during the 6 MWT. (b) Time needed to complete the TUG. (c) Gait speed during the 10 MWT. (d) Gait phases during the 6 MWT. (e) Cadence during the 10 MWT. (f) Step length in % of body height (%BH) during the 6 MWT. Bar plots show the mean and standard deviation from all the participants. The correlation was measured with the Spearman correlation coefficient ($\rho$). Stars (*) indicate statistically significant differences (p $<0.05$).

Figure 3

(a) Average swing pattern of the leg in the sagittal plane of each participant and device during the 6 MWT. The toe marker is shown with a different color (KAFO: light orange, ABLE: blue). (b) Representative knee flexion/extension angle of each participant during the gait cycle, and a bar plot showing the average and standard deviation from all the participants’ ROM. (c) Representative thigh segment angle in sagittal plane of each participant during the gait cycle, and a bar plot showing the average and standard deviation from all the participants’ ROM. The knee joint and thigh segment angles were measured during the 6 MWT. Stars (*) indicate statistically significant differences (p $<0.05$).

Figure 4

(a) Circumduction during the 6 MWT. Circumduction was defined as the maximum lateral difference of the ankle marker between stance and swing phase. Bars show the mean and standard deviation from all the participants. (b) The average ankle marker trajectory with respect to the PCOM in the transverse plane during the gait cycle for each participant. Black circles represent the PCOM, measured as the middle point between trochanters. (c) Box plot of the relative position of the PCOM with respect to the center of the leading foot at toe-off in both anteroposterior and mediolateral directions. Statistical analysis was done using the mean values using each device. (d) Trajectories of the PCOM (big circle) in the transverse plane during an average gait cycle for each participant. Dashed lines show the straight path to follow while walking. (e) Circuity index mean and standard deviation using each device. The circuity index is defined as the ratio of the total distance covered by the PCOM trajectory to the euclidean distance between its first and last position in the transverse plane during a gait cycle. Data were recorded during the 6 MWT. Stars (*) indicate statistically significant differences (p $<0.05$).

Figure 5

(a) Percentage of the peak oxygen uptake ($\%V{O}_{2peak}$) for each device, calculated by dividing the $V{O}_{2avg}$ by the $V{O}_{2peak}$ obtained from the maximal graded exercise during the preliminary assessment. The $V{O}_{2avg}$ is the average oxygen consumption over the last 2 min of the 6 MWT. The $V{O}_{2peak}$ is the maximum value recorded during the maximal graded exercise. Levels of physical intensity were defined as light (37–45 $\%V{O}_{2peak}$), moderate (46–63 $\%V{O}_{2peak}$), vigorous (64–91 $\%V{O}_{2peak}$), and maximal (>91 $\%V{O}_{2peak}$). (b) Walking efficiency during the 6 MWT for each device estimated by the Metabolic Cost of Transport (MCoT), according to the equation used by Martini et al.. Walking efficiency refers to the amount of oxygen consumed per 1 kg of body weight to walk a distance of 1 m.

Figure 6

(a) Scores of the three subscales of the PIADS questionnaire for each device. Scores range from − 3 (maximum negative impact) to zero (no perceived impact) to +3 (maximum positive impact). (b) First 8 items of the QUEST 2.0 questionnaire for each device. These items assess characteristics of the corresponding device in terms of the shown dimensions. Answers are categorized on a five-point scale that ranges from 1 (not satisfied at all) to 5 (very satisfied). Stars (*) indicate statistically significant differences (p $<0.05$).