Feasibility, safety, and functional outcomes using the neurological controlled Hybrid Assistive Limb exoskeleton (HAL®) following acute incomplete and complete spinal cord injury – Results of 50 patients

Abstract

Objective

Recent studies reveal that Hybrid Assistive Limb (HAL®) locomotion training in paraplegic patients suffering from chronic spinal cord injury (SCI) induces improvements in functional and ambulatory mobility. The purpose of this study was to determine the safety, feasibility, and functional effectiveness of HAL® locomotion training in the initial rehabilitation of acute SCI patients. This clinical trial represents the first systematic intervention worldwide for acute SCI patients using a neurologically controlled exoskeleton.

Design

Single center, prospective study.

Setting

BG University Hospital Bergmannsheil, Bochum, Germany.

Participants

Fifty acute SCI patients (14 women, 36 men).

Interventions

All participants received a daily (5 times/week) HAL® exoskeleton supported training for 12 weeks (mean amount of training sessions 60.4 ± 30.08).

Outcome Measures

Functional outcome for overground walking was monitored using the 10-m-walk test (10 MWT) combined with the WISCI II score, 6-minute-walk test (6 MWT) and the timed-up and go test (TUG test). Treadmill-related parameters (speed, distance and walking time) and the Lower Extremity Motor Score (LEMS) were recorded separately.

Results

Significant improvements were observed for HAL®-associated (walking time, distance and speed) and for functional outcomes (10 MWT, 6 MWT and TUG-test). WISCI-II-Score and the LEMS increased significantly compared with the status prior to training.

Conclusion

HAL® locomotion training is feasible and safe in the rehabilitation of acute SCI patients. The HAL® exoskeleton enables the patient to perform effective treadmill training and leads to improvements in functional and ambulatory mobility. However, spontaneous recovery vs training-related effects remain unclear and findings should not be extrapolated beyond the acute in-patient rehabilitation setting.

Trial registration: German Clinical Trials Register identifier: DRKS00010250..

Introduction

Traumatic spinal cord injury (SCI) occurs with an average yearly incidence varying from 40 to 80 cases per million inhabitants worldwide and is a wasteful and often disabling accident (1). SCIs are mostly of traumatic etiology, vehicle crashes (39.3%) and falls (31.8%) being the most frequent causes in the US. Recently, non-traumatic causes are increasing. Around 65% of these cases involve motor incomplete lesions (2).

Motor incomplete SCI shows an initial, naturally-occurring significant rate of spontaneous neurological recovery, promising an ambulation recovery rate of more than 75% (3). Nevertheless, the extent of neurological recovery depends on lesion level, patient's age, neurological impairment as assessed by the American Spinal Injury Association (ASIA) score, and early intensive and individually-tailored rehabilitation (3, 4). Neurological recovery is thought to be near completion within the first year post-injury, with the highest rate of improvement occurring within the first four months (5).

In the past 20 years, body weight supported treadmill training (BWSTT) has been established as a standard therapy to enhance motor function in incomplete SCI patients (6). BWSTT is composed of stepping movements on a motorized treadmill while a patient's weight is partially reduced by a harness system. As reported in several studies, BWSTT improves functional walking abilities, gait pattern and balance due to task-specific training. (7–9). Nevertheless, BWSTT often requires personal assistance by at least one physiotherapist to facilitate sufficient gait patterns. Driven gait orthoses (DGO) developed in the past decade, namely the Lokomat (Hocoma AG, Volketswil, Switzerland), enable a regular and intensive therapy which is essential for the maximum recovery of motor function in the early stages after a SCI (4, 8, 10–12). There have been no reports of significant functional outcome differences using a DGO compared to conventional locomotion therapy, despite the benefits of longer therapy duration and more physiological and reproducible walking patterns (13).

In recent years, exoskeletons like Ekso®, Rex®, Re-Walk®, and HAL® have been developed for SCI patient rehabilitation and/or the use as medical aids (14). Most of these systems are posture-controlled exoskeletons and enable standing and walking in patients with neurological gait disorders. However, the neurologically controlled Hybrid Assistive Limb exoskeleton (HAL®, Cyberdyne Inc., Japan) allows a voluntary driven range of motion. This is achieved via a neurological feedback system in which minimal bioelectrical signals are recorded and amplified from the lower extremities via cutaneous EMG electrodes (14–17). In a pilot study, Aach, Cruciger et al. reported on the feasibility and beneficial effects of HAL®-supported training in chronic SCI patients (15). In addition, Grasmücke et al. investigated the effect of 60 sessions of HAL® training on chronic SCI patients (17). Further, Sczesny-Kaiser et al. have hypothesized that HAL® training improves walking parameters and normalizes cortical excitability in the primary somatosensory cortex of chronic SCI patients (18).

According to the available literature, no systematic trial focusing on HAL®-locomotion training in acute SCI patients exists at this point. The purpose of this study was to examine whether HAL®-treadmill training is feasible and safe for use with functional incomplete SCI patients and capable of guiding the spontaneous recovery or inducing functional outcome improvements in these patients. We herein report on the neurological and functional results, as well as the training-related experiences following 60 sessions of HAL® exoskeleton therapy in acute SCI patients.

Material and methods

Patients

Fifty patients with acute incomplete and complete SCI (14 females, 36 males) participated in this prospective study. All patients were in the acute phase of SCI. In all patients, the time interval between SCI and the onset of HAL® training was less than one year (mean time 117.98 ± 95.82 days; range 4–327 days). All patients were classified prior to training according to the American Spinal Injury Association Impairment Scale (AIS). SCI lesions were located between C4 and L4 (AIS A-D). Sixteen subjects suffered from tetraplegia, and 34 from paraplegia. Three patients were classified as AIS grade A with no motor sensory function in the sacral segments S4/S5, but with zones of partial preservation (ZPP) below the lesion level. 30 patients were categorized as AIS grade C. Seventeen patients were classified as AIS grade D. No subject was classified as AIS grade B. Thirty-nine of the subjects had suffered traumatic spinal cord injuries. In four patients, massively prolapsed intervertebral discs caused incomplete paraplegia and in one patient, an epidural abscess with spondylodiscitis caused the SCI. In one patient, intramedullary cavernoma, and in another patient, anterior spinal artery syndrome caused the spinal cord lesion. One patient suffered from tuberculosis with osteolysis of the vertebra T7-T10. In one patient, a tumor infiltrated the spinal canal, one patient suffered cervical spinal canal stenosis and cervical myelopathy and in one patient, ischemic myelopathy caused the spinal cord lesion. Mean age ± SD at the time of enrollment was 43.88 ± 15.0 years (range, 18–72 years).

Inclusion criteria were as follows:

• Acute (time since injury less than 12 months) motor incomplete paraplegia or tetraplegia (AIS grade C or D) or motor complete paraplegia (AIS grade A) with ZPP

• Independent of injury level and ASIA classification, the included patients were required to demonstrate motor function of the hip and knee extensor and flexor muscle groups in order to be able to trigger and control the exoskeleton

• Age over 18 years

Exclusion criteria were defined as follows:

• Absence of residual motor functions

• Acute thrombosis or embolic diseases, epilepsy and severe heart insufficiency

• Severe limitation in the range of motion (ROM) of hip and knee joints due to lower-extremity contractures or distinctive spasticity

• Pressure sores

• Cognitive impairment

• Body weight > 100 kg

• Non-consolidated fractures

All participants gave written informed consent to participate in the trial. Consent to publish the anonymized data was confirmed. The ethical board committee of Bergmannsheil Hospital and the University of Bochum approved the intervention. The study was conducted according to the principles expressed in the Declaration of Helsinki (Table 1).

Table 1.

Patient characteristics.

Case	Sex	Age	Time since trauma (days)	Etiology	Level	AIS
1	f	18	319	Tuberculosis T7, T8, T9, T10	T10	C
2	m	43	103	#L1	T11	A, ZPP L3
3	m	72	37	Spinal stenosis C3/4, cervical spinal cord disease	C4	D
4	m	35	47	#C5	C4/C5	D
5	f	39	63	#C6, C7, spinal cord contusion	C6	C
6	m	53	129	#T4, T5, T6, T8	T6	C
7	m	19	48	#C5	C5	C
8	f	64	21	Discal prolapse 10/11	T12	C
9	m	28	136	Spinal cord contusion, discoligamental rupture C3/4 and 4/5	C5	D
10	m	54	57	#L1, L4, L5	L1	C
11	m	59	213	#T3, T4	T4	C
12	m	28	52	#L4	L4	D
13	m	31	42	#L1	L4	A, ZPP L4
14	m	45	33	#T 12	T12	C
15	m	38	148	#T12/ L1	T12	C
16	m	62	327	#T12	L1	C
17	m	62	110	Spinal cord contusion C2, C3, C4, discoligamental rupture, instability of segments C3, C4	C4	C
18	m	50	115	#C3	C4	C
19	m	50	129	#T12	T12	C
20	f	29	28	Cavernoma T1/T2	T4	D
21	f	52	36	Discoligamental rupture with spinal cord damage C4/5	C5	D
22	m	44	54	#T11, T12, L1, L2, L3, L4	T12	C
23	m	37	65	Discal prolapse T12/L1	T12	C
24	m	33	28	Discal prolapse C6/7	C7	C
25	f	68	199	Epidural abscess with discitis	T10	D
26	m	19	121	#T1, T 3, T 4	C7	C
27	m	60	84	#T1	C4	D
28	m	55	38	#T12, #L1	T12	A, ZPP L4
29	f	59	56	#T12, L1	L1	C
30	m	64	54	Central cord syndrome, discal prolapse C4-7	C5	D
31	f	56	73	#C3, 4, 5	C4	C
32	m	60	116	#C5	C5	C
33	m	56	259	#L1	L1	C
34	m	42	40	#T12	T12	D
35	m	55	275	Tumor infiltration of the cervical spinal canal	T11	D
36	m	52	46	B1 injury T5, 6, #T5-8,	T7	D
37	f	18	376	#C6, 7	C7	D
38	f	20	267	#L2	L2	C
39	m	45	42	#T3, T4	T8	C
40	m	27	304	#T11	T12	D
41	f	59	264	A. spinalis anterior syndrome T9/10	T12	C
42	m	36	45	#T7, 8	T7	C
43	m	30	185	#T12	L1	C
44	m	21	27	Stab wound of the spinal cord	T12	D
45	m	34	59	#T8	T10	D
46	f	50	51	#T11, T12	L1	C
47	f	48	94	Ischemic spinal cord disease T10	T10	C
48	m	21	129	#T11, T12	T12	C
49	m	46	278	#T12	T12	C
50	f	48	77	#C2	C4	D

m, male; f, female; #, fracture; ZPP, zones of partial preservation; C, cervical; T, thoracic; L, lumbal; S, sacral.

Intervention

All study participants underwent body weight supported training (BWSTT) on a treadmill (Woodway Inc.) five times per week for 60 training sessions (60.4 ± 30.08) using the HAL® exoskeleton (Cyberdyne Inc., Japan). The training was performed with individually adjustable weight bearing and walking speed. Training sessions included a 10-m-walk-test (10 MWT) before and after each session and regular physiotherapy, lasting approximately 90 min. Experienced physiotherapists who conducted the training were neither involved in study design nor data analysis. A physiotherapist and a medical doctor supervised the therapy. The study was performed between July 2013 and April 2019 in the BG University Hospital Bergmannsheil, Bochum.

The exoskeleton

The Hybrid Assistive Limb® (Cyberdyne Inc., Japan) is an exoskeleton consisting of a frame and actuating motors that is attached to patient's lower extremities. Volitional initiated contractions of the extensor and flexor muscles of hip and knee produce minimal bioelectrical signals which are detected by EMG electrodes and transmitted to a detachable controller. Hence, actuators compensate individually and independently for any absent/insufficient muscle strength to support the patient while treadmill training or overground walking (Fig. 1).

Figure 1.

HAL® – Hybrid Assistive Limb exoskeleton (Cyberdyne Inc.).

Outcome measures

In this study, we investigated the functional outcome, walking capabilities, and neurological status of acute SCI patients after 60 sessions of daily HAL® treadmill training. Testing was conducted prior to training and following 30 and 60 sessions of training. We used the assessment of the Lower Extremity Motor Score (LEMS) by manual muscle testing according to the AIS to determine muscle strength (19). Gait speed and cadence were recorded while performing the 10-meter-walk-test (20). To determine the required assistance for walking a 10 m distance the Walking Index for Spinal Cord Injury – II (WISCI II Score) was used (21). The last 10 MWT was conducted with the newly required assistance (WISCI II) by the end of HAL training. A declined result in 10 MWT could be attributed to the lessened required assistance for walking 10 meters. Gait endurance was measured by the 6-minute-walk-test (6 MWT). For this test, patients were instructed to walk using their personally required assistive devices at their preferred speed for a six-minute time period while covered distance (in meters) was logged. Patients could rest if they felt unable to continue the 6 MWT (22). If patients felt unable to walk for 6 min, the accomplished distance in a smaller amount of time was recorded. With the Timed-Up-And-Go-Test (TUG), the time required to complete multiple tasks including standing up from the wheelchair, walk 3 m, return to the wheelchair and sit down were recorded (23). All functional tests mentioned above were performed without the exoskeleton. Walking distance, speed, and walking time on the treadmill while using the exoskeleton were continuously recorded during training sessions. Walking time was limited to a maximum of approximately 30 min due to organizational reasons of maximal therapy time. For surveillance of the cardiovascular system, blood pressure and heart frequency were monitored throughout training. All adverse or severely adverse events were recorded.

Statistical analysis

Descriptive analysis of patient demographics (age, gender) and injury characteristics was performed using frequency distribution for categorical data and mean for continuous variables ± standard deviation (SD) to provide general information about the study population.

Data analysis was carried out using paired t-Test with the inner-subject-factor TIME (baseline- and 12w-condition). The differences were considered statistically significant at p < 0.05. Data were analyzed using SPSS (version 18.0, SPSS, Chicago, IL, USA).

Results

Treadmill associated results:

In all patients mean walking speed increased significantly (paired t-Test, p ≤ 0.000) from 0.9 ± 0.38 km/h (0.3–2.4 km/h) at baseline to 2.2 ± 0.7 km/h (0.7–4.2 km/h) at 60 sessions. Mean walking time increased significantly (paired t-Test, p ≤ 0.000) from 16.06 ± 5.68 min at baseline to 28.88 ± 6.43 min at 60 sessions. Mean ambulated distance increased significantly from 262 ± 181 m at baseline to 997 ± 411 m (paired t-Test, p ≤ 0.000) at 60 sessions (Fig. 2).

Figure 2.

Walking distance on Treadmill using HAL® at baseline and after 60 sessions.

Functional outcome

All patients demonstrated significant improvements in overground gait assessments following HAL® locomotion training. Mean walking time in the 10 MWT decreased significantly (p ≤ 0.000) from 64.1 ± 46.8 s at baseline to 26.5 ± 35.9 s at 60 sessions (Fig. 3). The number of steps decreased from 30.0 ± 9.2–20.36 ± 4.98 while step length increased from 36.03 ± 9.79 cm to 51.69 ± 11.33 cm. Accordingly, cadence (steps per minute) increased from 44.9 ± 25.35 steps/minute to 79.5 ± 33.59 steps/minute.

Figure 3.

10-Meter-Walk-Test at baseline and after 60 sessions.

Mean ambulated distance in the 6 MWT increased significantly from 135.58 ± 93.75 m at baseline to 233.33 ± 131.03 m at 60 sessions, while patients were walking 6 min without any break.

Similar results were found in the TUG test. The mean time required decreased from 56.25 ± 25.83 s at baseline to 28.76 ± 20.45 s at 60 sessions.

Statistically significant improvements in WISCI II scores were observed. The mean WISCI II score increased from 7.0 ± 4.8 at baseline to 13.3 ± 5.6 at 60 sessions (Fig. 4). Of the total 50 patients, 44 showed increased WISCI II scores. Seventeen of these patients were no longer dependent on walking aids following the intervention. The mean WISCI II improvement was 6.30 ± 5.05.

Figure 4.

Difference between the WISCI-II-Score before and after HAL® therapy.

The LEMS was assessed in all but one patient. From 49 patients, two patients LEMS remained the same (41 and 13) and in 47 patients the LEMS increased. Mean LEMS increased significantly (paired t-Test, p ≤ 0.000) from 27.47 ± 10.326 at baseline to 36.33 ± 9.955 at 60 sessions.

Feasibility and safety

No adverse nor severely adverse events such as falls, pressure ulcers, or other injuries occurred during the intervention. Skin redness in relation to the electrodes occurred in some patients. This redness resolved following removal of the electrodes. One patient developed a pressure ulcer of his left heel during a weekend homestay. Training was suspended for three weeks to guarantee the consolidation of his heel. The training intensity did not trigger any cardiac or respiratory-related problems. Especially in patients suffering from a lesion level T 5 or higher, orthostatic hypotension occurred during the first sessions but stabilized over the course of the training.

Discussion

The essential factor determining recovery from neurological and functional impairments after SCI is the extent of motor impairment. Functional and neurological outcomes rely inter alia on intensive and individually tailored rehabilitation in the early stages after a spinal cord injury. Aach et al. and Grasmücke et al. reported that HAL® exoskeleton training results in improved over ground walking and leads to a beneficial effect on ambulatory mobility in chronic SCI patients (15, 17). Hence, the primary aim of the present study was to identify whether HAL® exoskeleton locomotion training is safe and feasible, improves functional mobility, and increases motor functions in acute SCI patients. The results revealed significant improvements for over ground walking speed in the 10 MWT, increased gait endurance in the 6 MWT, and reduced time required for the TUG test. Speed and endurance improvements showed 61.5% faster walking and 49% greater distance covered following training. Patients significantly improved in 10 MWT time by 58.67%. This improvement needs to be considered knowing that the 10 MWT at the last training session was conducted using the newly required assistance (new WISCI II-Score). With the original required assistance, patients are likely to have completed the 10 MWT in a faster time. Our results of improved functional outcome are consistent with those of other studies using robot-assisted gait training as recently described in a meta-analysis by Fang et al. (24).

Further significant improvements in walking abilities as assessed by WISCI II scores were observed. Finally, there were significantly increased LEMS in all but two subjects. The safety and feasibility of the system was demonstrated.

The primary limitation of this study is the lack of a control group. Therefore, it is not yet possible to clarify the influence of spontaneous recovery on improved outcomes observed following HAL® therapy. Nevertheless, HAL® is by now the only fully neurologically-controlled exoskeleton. This leads to a permanent activation of the wearer.

A further limitation is the heterogeneity of the cohort. However, all patients demonstrated similar improvements independent of SCI etiology.

Due to rapid spontaneous recovery during the first months and up to one year after SCI a follow-up of more than a year is necessary (25). Further statistical limitations of the study include the relatively small number of patients (n = 50), as well as the mixture of complete and incomplete, paraplegic and quadriplegic SCI patients. On the other hand, the number of 50 acute SCI subjects in a mono centric trial according to an innovative therapy is unique.

The results demonstrated significant neurological improvements for acute SCI patients. Functional improvements after HAL®-assisted locomotion training in acute SCI patients included better overground walking without the exoskeleton and a reduction in required assistive devices. However, findings should not be extrapolated beyond the acute in-patient rehabilitation setting and it is necessary to continue research with large randomized and controlled trials to compare efficacy of HAL® assisted training with established conventional therapies in acute SCI treatment.

Even if we were able to demonstrate safety and feasibility of the system and functional improvements, accessibility of HAL® training is still restricted due to insurance situation of the patients which prevents a widespread use. However, in our clinic patients under the statutory accident insurance receive HAL® training on a regular base if suitable.

In conclusion, to the author’s knowledge the present study represents the first examination of neurological and functional results after voluntary driven HAL®-locomotion training in acute SCI patients.

Ethical approval

The ethical board committee of Bergmannsheil Hospital and the University of Bochum approved the intervention (Register No. 4733-13).

Disclaimer statements

Disclosure statement No potential conflict of interest was reported by the author(s).

Data availability statement The authors confirm that the data supporting the findings of this study are available within the article or its supplementary materials.

Funding Statement

This work was supported by the New Energy and Industrial Technology Development Organization, Japan (NEDO) [grant number not available] and a governmental grant (I&K-Gender-Study/European-Union and NRW/Germany under grant number 005-GW02-069B). The trial has been exclusively performed and supervised by the staff of BG University Hospital Bergmannsheil Bochum.