A Randomized and Controlled Crossover Study Investigating the Improvement of Walking and Posture Functions in Chronic Stroke Patients Using HAL Exoskeleton - The HALESTRO Study (HAL-Exoskeleton STROke Study)

Abstract

Background: The exoskeleton HAL (hybrid assistive limb) has proven to improve walking functions in spinal cord injury and chronic stroke patients when using it for body-weight supported treadmill training (BWSTT). Compared with other robotic devices, it offers the possibility to initiate movements actively. Previous studies on stroke patients did not compare HAL-BWSTT with conventional physiotherapy (CPT). Therefore, we performed a crossover clinical trial comparing CPT and HAL-BWSTT in chronic stroke patients with hemiparesis, the HALESTRO study. Our hypothesis was that HAL-training would have greater effects on walking and posture functions compared to a mixed-approach CPT. Methods: A total of 18 chronic stroke patients participated in this study. Treatment consisted of 30 CPT sessions and of 30 sessions of BWSTT with a double leg type HAL exoskeleton successively in a randomized, crossover study design. Primary outcome parameters were walking time and speed in 10-meter walk test (10MWT), time in timed-up-and-go test (TUG) and distance in 6-min walk test (6MWT). Secondary outcome parameters were the functional ambulatory categories (FAC) and the Berg-Balance Scale (BBS). Data were assessed at baseline, at crossover and at the end of the study, all without using and wearing HAL. Results: Our study demonstrate neither a significant difference in walking parameters nor in functional and balance parameters. When HAL-BWSTT was applied to naïve patients, it led to an improvement in walking parameters and in balance abilities. Pooling all data, we could show a significant effect in 10MWT, 6MWT, FAC and BBS, both therapies sequentially applied over 12 weeks. Thereby, FAC improve from dependent to independent category (3 to 4). One patient dropped out of the study due to intensive fatigue after each training session. Conclusion: HAL-BWSTT and mixed-approach CPT were effective therapies in chronic stroke patients. However, compared with CPT, HAL training with 30 sessions over 6 weeks was not more effective. The combination of both therapies led to an improvement of walking and balance functions. Robotic rehabilitation of walking disorders alone still lacks the proof of superiority in chronic stroke. Robotic treatment therapies and classical CPT rehabilitation concepts should be applied in an individualized therapy program.

Keywords: exoskeleton; hybrid assistive limb; locomotor training; physiotherapeutic approach; stroke rehabilitation.

FIGURE 1

Study design. Two-period, controlled crossover design. Each period included 30 therapy sessions. 1-week break was held between both periods. Before, at crossover and at the end of the study, assessments were done. HAL = hybrid assistive limb, Group 1 = HAL-CPT; Group 2 = CPT-HAL.

FIGURE 2

HAL exoskeleton and training setting. Left part of the figure shows HAL exoskeleton with its electronic actuators for hip and knee joints (“power units”), its battery pack and controller at the top and its EMG-electrodes and cables for detection of bioelectrical signals. Right part of the figure demonstrates the training setting with the treadmill, the body-weight support and the exoskeleton. Official picture of Cyberdyne Inc., the person was not subject of the published study and has been engaged and paid for commercials. Copyright Cyberdyne, Inc., published with kindly permission.

FIGURE 3

Walking performance parameter. Walking abilities at baseline, crossover and at the end of the study. Figure shows mean values for Group 1 (HAL-CPT) and Group 2 (CPT-HAL). Ten MWT = 10-meter walking test (A), 6 MWT = 6-min walking test (B), TUG = timed-up-and-go test (C). ^∗ indicated post hoc t-tests after significant effect for factor “time,” ^∗∗ indicated post hoc t-test after significant interaction between “time” and “Group” (see “Results” section for details). p-threshold < 0.017 (Bonferroni correction).

FIGURE 4

Functional outcome and Berg-Balance Scale. Figure shows mean values for Group 1 (HAL-CPT) and Group 2 (CPT-HAL). FAC = Functional ambulation categories (A), BBS = Berg-Balance-Scale (B). ^∗ indicated post hoc t-tests after significant effect for factor “time.” p-threshold < 0.017 (Bonferroni correction).