This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on http://www.researchprotocols.org, as well as this copyright and license information must be included.
In wheelchair users with a chronic spinal cord injury (WUSCI), prolonged nonactive sitting time and reduced physical activity—typically linked to this mode of mobility—contribute to the development or exacerbation of cardiorespiratory, musculoskeletal, and endocrine-metabolic health complications that are often linked to increased risks of chronic pain or psychological morbidity. Limited evidence suggests that engaging in a walking program with a wearable robotic exoskeleton may be a promising physical activity intervention to counter these detrimental health effects.
This study’s overall goals are as follows: (1) to determine the effects of a 16-week wearable robotic exoskeleton–assisted walking program on organic systems, functional capacities, and multifaceted psychosocial factors and (2) to determine self-reported satisfaction and perspectives with regard to the intervention and the device.
A total of 20 WUSCI, who have had their injuries for more than 18 months, will complete an overground wearable robotic exoskeleton–assisted walking program (34 sessions; 60 min/session) supervised by a physiotherapist over a 16-week period (one to three sessions/week). Data will be collected 1 month prior to the program, at the beginning, and at the end as well as 2 months after completing the program. Assessments will characterize sociodemographic characteristics; anthropometric parameters; sensorimotor impairments; pain; lower extremity range of motion and spasticity; wheelchair abilities; cardiorespiratory fitness; upper extremity strength; bone architecture and mineral density at the femur, tibia, and radius; total and regional body composition; health-related quality of life; and psychological health. Interviews and an online questionnaire will be conducted to measure users’ satisfaction levels and perspectives at the end of the program. Differences across measurement times will be verified using appropriate parametric or nonparametric analyses of variance for repeated measures.
This study is currently underway with active recruitment in Montréal, Québec, Canada. Results are expected in the spring of 2021.
The results from this study will be essential to guide the development, implementation, and evaluation of future evidence-based wearable robotic exoskeleton–assisted walking programs offered in the community, and to initiate a reflection regarding the use of wearable robotic exoskeletons during initial rehabilitation following a spinal cord injury.
ClinicalTrials.gov NCT03989752; https://clinicaltrials.gov/ct2/show/NCT03989752
DERR1-10.2196/19251
Approximately 100,000 Canadians are currently living with a
Given the increased life expectancy owing to improvements in medical treatment, along with the growing population of individuals with SCI, there have been recent calls to direct additional attention to the cascade of cardiorespiratory, musculoskeletal, and endocrine-metabolic health problems faced by this population and to interventions targeting modifiable factors linked to these problems [
Gravity-derived high-standing loads, as well as impacts resulting even from low walking speeds [
Two to three sessions per week of regular structured exercise at moderate-to-vigorous intensity for at least 20 minutes, plus upper body strength exercise (ie, three sets of 10 repetitions at 50%-80% of the one-repetition maximum for large muscle groups), improve cardiorespiratory and endocrine-metabolic health [
Commercially available wearable robotic exoskeletons, an assistive technology allowing WUSCI to stand and walk overground (see
Wearable robotic exoskeleton for sit-stand transitions and overground walking manufactured by Ekso Bionics.
The
Question 1: Does a 16-week walking program with the wearable robotic exoskeleton induce beneficial changes on musculoskeletal, cardiorespiratory, and endocrine-metabolic health; wheelchair-related functional skills and mobility; and psychosocial outcomes? Hypothesis 1: It is hypothesized that beneficial effects observed during the postintervention and retention measurement times will significantly and meaningfully exceed any changes observed during the control and preintervention measurement times (ie, T0 [control measurement time] vs T1 [preintervention measurement time] vs T2 [postintervention measurement time] vs T3 [retention measurement time]).
Question 2: What personal factors best determine and predict the beneficial effects of the walking program with the wearable robotic exoskeleton? Hypothesis 2: It is hypothesized that the individuals with the highest level of SCI and the longest time since the SCI (ie, possibly the best determinants and predictors) will be those who respond best to the walking program.
Question 3: What program attributes best determine and predict the beneficial effects of the walking program with the wearable robotic exoskeleton? Hypothesis 3: It is hypothesized that the total number of steps taken will be the best determinant and predictor of the measured changes.
Question 4: What are the participants’ satisfaction levels with the walking program and the wearable robotic exoskeleton itself, and what are the expectations regarding its future use in the context of a home- or community-based adapted physical activity program? Hypothesis 4: It is hypothesized that WUSCI will (1) express high levels of satisfaction with the walking program using the wearable robotic exoskeleton and with the wearable robotic exoskeleton itself and (2) report on how they envision its future in the context of home- or community-based use to shape the development of an adapted physical activity program in the future.
A prospective, longitudinal, self-controlled interventional study with multiple discrete measurement times will be used to assess outcomes at baseline (ie, preintervention phase), during the intervention, and thereafter (ie, retention phases) (see
Summary of the design of the study along with the different assessment times. T0: control measurement time; T1: preintervention measurement time; T2: postintervention measurement time; T3: retention measurement time.
We aim to recruit a nonprobabilistic consecutive sample of 20 long-term WUSCI. The inclusion and exclusion criteria are listed in
Participant-specific inclusion criteria:
Adults (≥18 years old)
Chronic complete or incomplete traumatic or nontraumatic spinal cord injury (SCI) at least 18 months before enrollment
Long-term manual wheelchair use as primary means for in-house and community mobility (ie, nonambulatory)
Understand and communicate in English or French
Reside or will arrange for temporary housing in the community within 75 km from the main research site
Participant-specific exclusion criteria:
Other neurological impairments aside from those linked to the SCI (eg, multiple sclerosis)
Concomitant or secondary musculoskeletal impairments (eg, hip heterotopic ossification)
History of lower extremity fracture within the past year
Unstable cardiovascular or autonomic system
Renal insufficiency
Pregnancy
Any other conditions that may preclude lower extremity weight-bearing, walking, or exercise tolerance in the wearable robotic exoskeleton
Exoskeleton-specific inclusion criteria:
Body mass: ≤100 kg
Height: 1.52-1.93 m
Pelvis width: 30-46 cm
Thigh length: 51.0-61.4 cm
Lower leg length: 48.0-63.4 cm
Exoskeleton-specific exclusion criteria:
Inability to sit with hips and knees at ≥90° flexion
Lower extremity passive range of motion limitations (hip flexion contracture ≥5°, knee flexion contracture ≥10°, and ankle dorsiflexion ≤–5° with knee fully extended)
Moderate-to-severe lower extremity spasticity (score of >3 on the Modified Ashworth Scale)
Length discrepancy (≥1.3 cm or ≥1.9 cm at the thigh or lower leg segment, respectively)
Skin integrity issues preventing wear of the wearable robotic exoskeleton
An Ekso GT (Ekso Bionics) wearable robotic exoskeleton, which has been approved by Health Canada, is used in this study (see
Progression of the number of training sessions per week during the 16-week walking program.
Total hip areal bone mineral density (aBMD), determined with dual-energy x-ray absorptiometry (DXA) scans performed at T1, is used to assign each participant to one of three training regimes based on lower extremity fracture risks [
To be considered as having successfully completed the program, at least 75% of the training sessions (ie, 26/34) need to have been completed. To this effect, to assure an optimal attendance rate similar to the one reached during the feasibility study (ie, attendance rate was 97.7%) [
All outcomes reflecting the potential impacts of the intervention based on the logic model (see
Project-specific logic model highlighting the relationships between the different domains of interest and related outcome measures. L/E: lower extremity; SCI: spinal cord injury; U/E: upper extremity; WRE: wearable robotic exoskeleton.
Summary of outcomes.
Outcomes | Measurement timesa | |||||
|
T0 | T1 | T2 | T3 | ||
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
Sociodemographic characteristics (age, sex, etc) | ✓ |
|
|
|
|
|
Neurological impairment (American Spinal Injury Association Impairment Scale) | ✓ |
|
|
|
|
|
Anthropometric parameters (weight and height) | ✓ |
|
|
|
|
|
Resting heart rate and blood pressure | ✓ | ✓ | ✓ | ✓ |
|
|
Pain (International SCI [spinal cord injury] Pain Basic Dataset version 2.0) | ✓ | ✓ | ✓ | ✓ |
|
|
Passive range of motion at the ankle, knee, and hip joints (two-axis goniometer) | ✓ | ✓ | ✓ | ✓ |
|
|
Spasticity (Modified Ashworth Scale) | ✓ | ✓ | ✓ | ✓ |
|
|
|
|
|
|
|
|
|
20-meter wheelchair propulsion test (natural and maximal speeds) | ✓ | ✓ | ✓ | ✓ |
|
|
Slalom test | ✓ | ✓ | ✓ | ✓ |
|
|
6-minute manual wheelchair propulsion test | ✓ | ✓ | ✓ | ✓ |
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
Dual-energy x-ray absorptiometry (hip and lumbar vertebrae) | ✓ | ✓ | ✓ | ✓ |
|
|
Peripheral quantitative computed tomography (proximal tibia, distal femur, and proximal radius) | ✓ | ✓ | ✓ | ✓ |
|
|
|
|
|
|
|
|
|
Dual-energy x-ray absorptiometry (total body) | ✓ | ✓ | ✓ | ✓ |
|
|
|
|
|
|
|
|
|
Peripheral quantitative computed tomography (intramuscular fat infiltration) | ✓ | ✓ | ✓ | ✓ |
|
|
|
|
|
|
|
|
|
Bone turnover (serum procollagen type 1 N-terminal peptide, osteocalcin, C-terminal cross-linking telopeptide, and 25-hydroxyvitamin D) | ✓ | ✓ | ✓ | ✓ |
|
|
Glycemia (fasting glucose, insulin, and glycosylated hemoglobin) | ✓ | ✓ | ✓ | ✓ |
|
|
Insulin resistance (homeostatic model assessment) | ✓ | ✓ | ✓ | ✓ |
|
|
Lipids (total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, and apolipoprotein B) | ✓ | ✓ | ✓ | ✓ |
|
|
Inflammation (C-reactive protein, tumor necrosis factor alpha, interleuken-6, and interleuken-10) | ✓ | ✓ | ✓ | ✓ |
|
|
|
|
|
|
|
|
|
Respiratory gas analysis during 6-minute manual wheelchair propulsion test |
|
✓ | ✓ |
|
|
|
Total distance travelled during the 6-minute manual wheelchair propulsion test | ✓ | ✓ | ✓ | ✓ |
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
World Health Organization Quality of Life assessment | ✓ | ✓ | ✓ | ✓ |
|
|
Beck Depression Inventory | ✓ | ✓ | ✓ | ✓ |
|
|
Beck Anxiety Inventory | ✓ | ✓ | ✓ | ✓ |
|
|
Psychological General Well-Being Index | ✓ | ✓ | ✓ | ✓ |
|
|
|
|
|
|
|
|
|
Updated version of the Montreal Walking Exoskeleton Satisfaction and Perspectives Questionnaire |
|
|
✓ |
|
|
|
Semistructured interview |
|
|
✓ |
|
aMeasurement times: T0 (control measurement time), T1 (preintervention measurement time), T2 (postintervention measurement time), and T3 (retention measurement time).
Assessments are completed at the different assessment times to collect outcomes characterizing the following:
Sociodemographic characteristics (eg, age; sex; time since injury; history of fragility fracture; medications, including opioid analgesia, benzodiazepines, or unfractionated heparin; current smoking status; and alcohol intake) [
Neurological impairment (eg, American Spinal Injury Association Impairment Scale for neurological level, motor and sensory scores, and severity) [
Anthropometric parameters (eg, weight and height).
Resting heart rate and systolic and diastolic blood pressure using an electronic sphygmomanometer machine [
Pain using the
Passive range of motion at the ankle, knee, and hip joints with a two-axis goniometer (ie, contracture).
Upper extremity muscle strength (ie, pushing and pulling strength with a wheelchair wheel attached to an instrumented dynamometer and handgrip strength with a handheld dynamometer).
Lower extremity spasticity using the
Wheelchair abilities are assessed using the following performance-based wheelchair propulsion tests: (1) 20-meter wheelchair propulsion test (natural and maximal speeds), (2) slalom test, and (3) 6-minute manual wheelchair propulsion test [
A DXA system (Lunar Prodigy, GE Healthcare) is used to calculate aBMD at the hip, femoral neck, and the first to the fourth lumbar vertebrae [
In addition, a peripheral quantitative computed tomography (pQCT) system (XCT 3000, Stratec Biomedical Systems) is used to characterize the volumetric bone mineral density (vBMD) and the microarchitecture parameters of trabecular and cortical bones at various imaging sites: 66% of the tibia, 25% of the femur, and 66% of the radius. These sites were chosen to maximize muscle circumference in each scan [
Whole-body scans obtained with the DXA system are used to quantify total and regional (ie
Cross-sectional images of the femur, tibia, and radius captured with the pQCT system are also used to measure the muscle size (ie, cross-sectional area) and intramuscular fat infiltration (ie, muscle density) using the same validated open source image analysis software [
Fasting blood samples (ie, >8-hour fast) are used to quantify bone turnover biomarkers (ie, serum procollagen type 1 N-terminal peptide, osteocalcin, C-terminal cross-linking telopeptide, and 25-hydroxyvitamin D), glycemic biomarkers (ie, fasting glucose, insulin, and glycosylated hemoglobin), insulin resistance biomarkers (ie, homeostatic model assessment), lipid biomarkers (ie, total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, and apolipoprotein B), and inflammatory biomarkers (ie, C-reactive protein, tumor necrosis factor alpha, interleuken-6, and interleuken-10).
At T1 and T2, participants complete the 6-minute manual wheelchair propulsion test wearing a gas analyzer system (COSMED K4b2, COSMED srl). This portable system incorporates a sealed face mask placed over the mouth and nose and anchored around the head, a telemetric stationary O2 and CO2 gas analyzing unit, and a battery harnessed to the anterior and posterior thorax. This system is calibrated before each test as recommended by the manufacturer. For the other two measurement times (ie, T0 and T3), the total distance travelled during the 6-minute manual wheelchair propulsion test is used as a surrogate measure of cardiorespiratory fitness since it has been found to strongly correlate and agree with the maximal arm-crank test (
For the psychosocial outcomes, health-related quality of life is measured with the short version of the
For participant satisfaction, an updated version of the
For participants’ perspectives, a 20-30-minute semistructured interview is conducted over the phone to capture their general experience when participating in the wearable robotic exoskeleton–assisted training program. The interviews also serve as a platform for documenting participants’ perspectives on the future of wearable robotic exoskeleton technology. To do so, various themes are discussed: potential benefits and recommendations for wearable robotic exoskeleton–assisted walking during the acute and subacute phases following spinal cord injury (eg, timing and conditions) or in a clinical setting during the chronic phase; opportunities for improvement (eg, functionality and structural aspects of the wearable robotic exoskeleton); and recommendations for a future home- or community-based wearable robotic exoskeleton–assisted walking program (eg, stairs, different surfaces, donning and doffing the wearable robotic exoskeleton without assistance, and operating the wearable robotic exoskeleton without assistance). The research professional who conducts these interviews has never met the participants and is not a member of the research team. All interviews are recorded to later enable verbatim transcription.
The sample size estimate was based on a comparison using the variability of the absolute change (ie, mean ±SD) in both body composition and bone mineral density (ie, main outcomes) measured pre- and postintervention in our preliminary study [
Descriptive statistics (eg, mean, SD, and 95% confidence interval) will be calculated for data summarizing sociodemographic characteristics as well as clinical and laboratory outcomes collected at the different measurement times. The normality of all data distributions and the absence of outliers will be verified via the Shapiro-Wilk test of normality and the absence of studentized residuals greater than ±3 SDs, respectively. Whenever applicable, the level of significance will be set at
For Hypothesis 1, one-way analysis of variance for repeated measures (ie, normally distributed continuous data) or Freidman tests (ie, non-normally distributed continuous or categorical data) with planned comparisons based on the hypothesis and Bonferroni correction will be conducted to detect significant time effects, with a special interest for the preintervention (T0 vs T1), intervention (T1 vs T2), and retention (T2 vs T3) phases. In accordance with the principles of a classic intention-to-treat approach, all participants will be included in the final analyses, regardless of withdrawal, compliance, or unintentional missing data. For missing data, imputation of the mean value for the specific group at the specific assessment time will be used. Per-protocol exploratory analyses will also be performed comparing outcomes for those with walking program compliances of greater than 75% to examine maximum treatment efficacy.
For Hypotheses 2 and 3, Pearson or Spearman correlation coefficients will investigate the strength and direction of the relationships between the overall observed changes ([T1–T2] / T1 × 100) for each personal factor (ie, dependent variable) and the program characteristics (ie, independent variable). Independent variables having reached a critical threshold (
For Hypothesis 4
This study was recently initiated at the
This project innovates by being among the first studies to comprehensively, prospectively, and longitudinally investigate the effects of a wearable robotic exoskeleton–assisted walking program among long-term WUSCI who have a very poor prognosis for walking recovery [
Many stakeholders may benefit from this interventional study. For WUSCI who have no or very limited walking ability, the walking program with the wearable robotic exoskeleton is not expected to have any reversal effect on their walking capacity without this novel mobility assistive technology. However, this project is relevant since it will generate the first evidence of the anticipated cardiorespiratory, musculoskeletal, and endocrine-metabolic health adaptations upon completion of a walking program with a wearable robotic exoskeleton. For the first time, the extent to which these adaptations translate into beneficial effects on functional capacity will also be verified, as will their effects on health-related quality of life and psychological health. This includes the psychological well-being domain, which was not specifically measured during the feasibility study but was mentioned by the majority of participants. Given the fact that the population of WUSCI continues to grow and that they now live longer, these potential beneficial effects are further warranted. The caregiver burden and the potentially costly long-term expenditures associated with adverse health events may also decrease. For rehabilitation professionals, the proposed project is relevant since strengthened evidence regarding the effects of the walking program and the characteristics of the best responders will be generated and will inform clinical decision-making processes or the development of a clinical algorithm for referring individuals with SCIs to a walking program. For rehabilitation program administrators and policy makers, the proposed study is relevant since the evidence generated may further confirm the need for publicly funded clinical and technological infrastructures to create structured programs incorporating walking technologies, such as the wearable robotic exoskeleton, and outcome measures into rehabilitation or adapted physical activity centers. Both program administrators and policy makers will need to work collaboratively and cohesively to develop creative solutions to address this current service gap and engage in transformative improvements. For the research community, this project provides a unique opportunity to create a strong multidisciplinary team of well-established scientists with diverse and complementary academic training as well as clinical and fundamental research expertise. For manufacturers with an interest in wearable robotic exoskeletons, among others, this project is relevant since the input from powered exoskeleton end users (ie, WUSCI) will become available and may enrich the continuous quality improvement process. This process is imperative to further support and accelerate the development of wearable robotic exoskeletons and to reach key commercialization milestones for the technology to become personalized and accessible for WUSCI interested in home or community use (ie, neuroprosthesis) in the next decade.
A few potential challenges merit attention:
Some potential participants will have insufficient passive range of motion at the lower extremities to engage with the project. These participants will be provided with a 4- to 6-week home-based stretching program, will be reassessed, and may become eligible later.
Female WUSCI may be underrepresented. Efforts will be made for the sample to be representative of the SCI population and to have women make up 20% of the sample. However, since the minority of individuals affected by SCIs are female and the sample size is limited (n=20), it is unlikely that statistical analysis by subgroups (ie, male vs female) will be feasible in order to account for potential sex and gender differences. Nonetheless, descriptive statistics will present results separately for women and men whenever indicated.
A small number of participants may demonstrate vitamin D deficiencies [
Some participants may concurrently engage in extraneous physical activity or may seek cointerventions during the project. Participants will be asked to maintain their customary level of physical activity during the project and to avoid engaging in new cointerventions. Any unintended intervention (ie, contamination or cointervention) that may influence the results will be documented and its effect carefully verified by the research team and possibly considered as a dichotomous variable (ie, present vs absent).
Some participants may experience some lower extremity neurorecovery. In the event a participant was to experience neurorecovery (ie, lower extremity motor score of ≥20 on the American Spinal Injury Association Impairment Scale), he or she would be withdrawn from the project and referred for a comprehensive neurological assessment and to an advanced locomotor training program.
areal bone mineral density
Centre de recherche interdisciplinaire en réadaptation du Montréal métropolitain
dual-energy x-ray absorptiometry
Montreal Walking Exoskeleton Satisfaction and Perspectives Questionnaire
peripheral quantitative computed tomography
spinal cord injury
control measurement time
preintervention measurement time
postintervention measurement time
retention measurement time
volumetric bone mineral density
wheelchair users with a chronic spinal cord injury
This study is funded by the
None declared.