Testing neuromuscular feedback therapy in patients with chronic spinal cord injury with residual motor function in the lower extremities compared to ‘conventional’ therapy: protocol for a multicentre randomised controlled trial

Abstract

Study design

Multicentre randomised controlled trial.

Setting

Eight spinal cord injury (SCI) centres and 15 rehabilitation centres in Germany.

Objectives

To demonstrate the superiority of neuromuscular feedback therapy using an exoskeleton compared to conventional therapy in terms of improving walking ability in patients with chronic SCI.

Methods

A total of 180 patients with chronic complete or incomplete SCI and residual motor function in the lower extremities are to be included. Baseline examination will be carried out at the recruiting SCI centres, followed by 1:1 randomisation at the rehabilitation centre allocating patients to the control group (conventional therapy) or the intervention group (Hybrid Assistive Limb therapy). Both interventions take place to the same extent (5×/week, 60 min) over a period of 12 weeks at the rehabilitation centres. After the 12-week intervention phase, the primary (therapy response defined as relevant improvement in at least one of the following: 6 min walk test, 10 m walk test and Walking Index for Spinal Cord Injury II score) and secondary endpoints (quality of life, pain, bladder and bowel function) are recorded by blinded observers at the SCI centres or via patient app. The effects of the intervention will be evaluated in the middle and at the end of the 6-month follow-up period.

Ethics and dissemination

The study is approved by the ethics committee of the Medical Association of Westphalia-Lippe (register no. 2025–2014 f-S). The participants were informed about the aims and procedures of the study and gave their written consent. This protocol follows the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) statement. The study results will be published in accordance with the Consolidated Standards of Reporting Trials (CONSORT), regardless of whether the results are positive or negative. The datasets belong to the Joint Federal Committee of Germany and are available there on request.

Trial registration number

DRKS 00035487.

Strengths and limitations of this study.

• High external validity: the multicentre design increases generalisability by including diverse clinics, therapists and patient populations.

• Good internal validity through randomisation: random allocation minimises selection bias and helps ensure comparable groups.

• Active control group improves fairness: both groups receive an intervention, allowing for a fair comparison, reducing ethical concerns and enabling an accurate estimation of the true effect of the Hybrid Assistive Limb.

• Blinded observer design limits only detection bias: since patients and therapists know the group, there is a risk of behaviour-related biases, for example, performance bias.

• Potential centre effects and variability: differences in protocol adherence, therapist experience and device handling across centres can dilute or distort treatment effects.

Introduction

Background and rationale

After the onset of acute spinal cord injury (SCI), the treatment of choice for incomplete paralysis and complete paralysis with zones of partial preservation is neurological rehabilitation treatment, including locomotion training.¹ This involves training the patient on a treadmill with partial removal of body weight. Intensive training over several weeks can increase the walking distance and improve the gait pattern.^2–4

The treatment of neurological deficits after SCI is still a major challenge. Although patients with SCI, in particular, have been able to achieve a high degree of autonomy through optimal wheelchair adaptation and appropriate rehabilitative therapies, a significant functional improvement in walking has not yet been achieved, depending on the severity of the injury. The so-called exoskeletons have recently been developed for patients with complete SCI, in particular, which are able to move paralysed patients in an upright position with a fixed gait cycle. However, there is no ‘communication’ in the sense of biofeedback with these systems. Extensive clinical studies on the effectiveness of targeted gait training in patients with walking disabilities have shown a positive influence on the musculoskeletal system with improvement in uprightness and walking ability, overall motor function, coordination and balance.^5–9

The Hybrid Assistive Limb (HAL; Cyberdyne Inc, Tsukuba, Japan) is also an exoskeleton. This can be operated in various programmes. In addition to the previously mentioned function, minimal bioelectric signals from the partially paralysed muscles can be recorded and supported by the motors on the hip and knee joints as required.¹⁰ This allows the patient to initiate a gait cycle at will. There is a link between the patient’s voluntary intention and the functional support of the exoskeleton. Accordingly, biofeedback training is also carried out with each training session. The device can be used by patients who have voluntary muscle functions in the lower extremities that are sufficient to control the device, regardless of the neurological classification according to American Spinal Injury Association Impairment Scale (AIS) as a component of the International Standards for Neurological Classification of SCI (ISNCSCI).

Previous studies have shown functional improvements through the use of the HAL exoskeleton in patients with neurological deficits.^11–17 In contrast to previous robot-assisted gait orthoses, patients can actively control the movement elements of the HAL via minimal residual muscle potentials of the leg muscles. The active cooperation of the patient results in a biofeedback loop.¹⁸ This approach has not yet been possible with any other available system. Compared with conventional treadmill therapy, due to the motor support, training with the HAL offers the potential to complete longer walking times in the same time interval, to take more steps and to train more physiologically.¹⁹ An increased rehabilitation potential with the possibility of functional improvement is expected. A recently published, randomised pilot study with a small number of cases also showed the potential of voluntary controlled movement therapy using an exoskeleton (Ekso GT, Ekso Bionics Holdings, San Rafael, California, United States) compared with locomotion therapy in terms of improving walking ability.²⁰ However, the mode of operation of the exoskeleton used in this study differs from that of the HAL system and does not use neuromuscular feedback therapy.

There are currently no comparative studies of neuromuscular feedback therapy versus conventional therapy in accordance with the therapeutic products directive. This study aims to close this knowledge gap.

Objectives

The aim of the study is to demonstrate the superiority of neuromuscular feedback therapy using the HAL exoskeleton compared with conventional therapy in accordance with the therapeutic products directive in terms of improving walking ability in patients with chronic SCI and residual motor function in the lower extremities (>1 year after the onset of SCI).

Methods

Ethics and dissemination

The study is approved by the ethics committee of the Medical Association of Westphalia-Lippe (register no. 2025–2014 f-S). The participants were informed about the aims and procedures of the study and gave their written consent. The participant consent form can be found in the online supplemental material 1. Since this study is conducted exclusively in Germany, this document is available only in German. This protocol is written following the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) statement. The SPIRIT checklist can be found in the online supplemental material 2.

Patient and public involvement

Patients could not be involved in designing the study or the interventions because this is a commissioned study by the federal government, which had already set specific, unalterable requirements for the study design.

Trial design

Multicentre randomised controlled trial with a parallel group design and balanced 1:1 randomisation using covariate imbalance minimisation.

Study setting

Eight SCI centres and 15 rehabilitation centres in Germany are planned. A list of study sites can be found in the clinical register. The SCI centres are responsible for patient recruitment, final study enrolment, and the blinded assessment of primary and secondary endpoints. After the baseline visit, patients are assigned to a rehabilitation centre of their choice or one that is available. The rehabilitation centres are responsible for randomisation and the execution of the respective therapies. An interim visit will also take place at the rehabilitation centre. The study is scheduled to begin in the third quarter of 2026 and is expected to conclude in the third quarter of 2028.

Eligibility criteria

Eligibility criteria for participants

The inclusion and exclusion criteria are checked in the recruiting SCI centres. Specifically, the following inclusion and exclusion criteria are defined.

Inclusion criteria:

1. Chronic, incomplete cervical, thoracic or lumbar SCI or complete SCI with residual motor function in the lower extremities.

2. More than 1 year after the onset of SCI.

3. Treatment at the expense of the public health insurance.

4. Height between 150 cm and 190 cm, weight between 40 kg and 100 kg.

5. Age ≥18 years and ≤75 years.

6. No further improvement in walking ability to be expected with conventional therapy in line with therapeutic products directive.

Exclusion criteria:

1. Contractures.

2. Massive spinal spasticity.

3. Cardiac devices or other electronic implants that may be affected by the use of the exoskeleton.

4. Pre-existing conditions that lead to an increased risk for the patient, such as:

   • Severe osteoporosis or condition following spontaneous fractures, incompletely consolidated fractures.

   • Coronary heart disease, pulmonary embolism in the past.

   • Neurological diseases such as Parkinson’s disease, epilepsy and severe dementia.

   • Psychiatric illnesses such as bipolar disorder, anxiety disorder, alcohol or drug addiction.

   • Decubital ulceration.

   • More than five training sessions with the HAL in the past.

   • Pregnancy and breastfeeding.

5. Lack of language comprehension or lack of capacity to consent.

The above list is not to be considered exhaustive; the final assessment is at the discretion of the including physicians. The selected inclusion/exclusion criteria were defined in such a way that the results can be transferred to the target population.

Eligibility criteria for sites and for individuals delivering the interventions

All therapists and physicians in the rehabilitation centres who are responsible for carrying out the study interventions must complete 3 days of intensive training in the use of the HAL (basic settings, customisations and training options) provided by Cyberdyne Care Robotics GmbH (Bochum, Germany). The conventional therapy is also carried out by specially trained therapists. Before starting, they receive training from qualified physiotherapists who have experience of working with people with SCI. A therapy guide and sample therapy plans are provided, and there is the option of contacting the trainers directly if there are any questions. Training sessions on the proper implementation and documentation of all primary and secondary endpoints will also be conducted as part of an initiation training programme. Rehabilitation centres must be equipped with a treadmill with overhead protection, a ramp to enable wheelchair-bound patients to reach the treadmill and a therapy couch.

Recruitment

Recruitment will primarily take place via the participating SCI centres, which will write to their existing patients from the past 10 years who fulfil the inclusion criteria with information about the study (cover letter and flyer). Around 1000 patients are to be reached in this way. Further recruitment will take place via self-help groups, flyers in doctors’ surgeries and press releases. If patients are interested, they can either contact a participating SCI centre of their choice directly or register for the study via the study website using a web application (Research Electronic Data Capture (REDCap, Vanderbilt University, Nashville, Tennessee, United States). In both cases, pre-screening is carried out using a questionnaire.

The final study inclusion and signing of the patient consent form takes place during the baseline visit in the SCI centres by experienced physicians.

Sample size determination

A reanalysis of the data from Grasmücke et al¹¹ revealed response rates for the individual response criteria (6 min walk test (6MWT), 10 m walk test (10MWT) and Walking Index for Spinal Cord Injury II (WISCI II) score) ranging from 45% to 55%. The combined response criterion showed a response rate of 82%. Therefore, a response rate of 80% under HAL therapy is assumed for sample size planning. Based on data from various studies, it can be expected that response rates under the control intervention for the two response criteria related to walking speed range between 20% and 35%.^20–22 Since the application of the combined response criterion suggests a higher proportion achieving a therapeutic response, a value of 55% is assumed for the control group in sample size planning. A sample size calculation using a two-sided Fisher’s exact test at a significance level of 5% indicated that 79 patients per group are required to achieve a power of 90%. With this sample size, the power to detect a difference of 25% remains above 80%, regardless of the response rate in the control group. Since the number of perceived sessions varied in both the HAL therapy study and the control therapy, the assumed response rates should provide a good approximation of the response rate even under potential non-compliance. Assuming a drop-out rate of approximately 10%, a total of 180 patients will be included in the study.

Randomisation

A balanced randomisation will be conducted using minimisation of covariate imbalance.^23–25 The following covariates will be considered: centre, WISCI II ≤5 versus > 5 points, age ≤50 years versus > 50 years, and spasticity. Since no studies on predictive factors under the control or HAL intervention are currently available, the selection of factors to be included was based on clinical judgement. Additional factors were excluded due to a suspected strong correlation with the selected ones, such as lesion level with spasticity and AIS classification.

Randomisation will be carried out by the local study leader in the rehabilitation centres shortly before the first appointment of patients using an electronic data capture system. For data capturing and data management of this clinical trial, a web-based validated software (AlcedisTRIAL, Alcedis GmbH, Gießen, Germany) will be used. The development and setup is done by applying Good Automated Manufacturing Practice standards. This ensures compliance with allocation concealment and maintains blinding of the assessors who will evaluate the endpoints in the SCI centres. Blinding of patients and therapists is not feasible. Patients will be instructed not to disclose the type of treatment they received during endpoint assessment in the SCI centres. The analysing biostatistician will be blinded to the group allocations.

Interventions

The planned intervention is neuromuscular feedback therapy (with the HAL exoskeleton). Before the intervention, the patient’s free mobility of the large joints is checked and any spasticity is reduced by brief passive stretching of the muscles. Electrodes are then attached to the patient’s flexor and extensor muscles of the hips (rectus femoris and gluteus maximus) and knees (biceps femoris or semitendinosus and the vastus lateralis or vastus medialis). The selection of electrode placement is based on the HAL user manual. Two self-adhesive disposable electrodes are used for each muscle group. These electrodes are later connected to the HAL via cables and form the basis for neuromuscular feedback training. The patient is then fitted with a patient harness, which can be connected to a safety system on the ceiling above the treadmill. This serves on the one hand to relieve any necessary weight when walking and on the other hand to ensure the patient’s safety. The patient’s feet are placed in the pressure-soled shoes of the exoskeleton and the entire system is closely connected to the patient via straps and rails. The HAL is also connected to the safety system. Finally, the cables of the electrodes are connected to the HAL and the system is started.

The intervention is planned as an individual intervention in personal contact on site. Doctors delegate the provision of feedback therapy to non-medical staff (physiotherapists, occupational therapists and sports scientists) but continue to bear the main responsibility for its implementation. One patient at a time is supervised by a therapist trained in the use of the HAL system for this indication.

The intervention is planned along the lines of the previous studies.^11–19 A therapy algorithm of five training sessions per week (Monday to Friday) over a period of 12 weeks, and thus 60 training sessions, is intended. Training beyond the 12-week period did not show any relevant further improvement.¹² Each training session comprises up to 30 min of walking time, excluding preparation and follow-up. It is estimated that a training session including stretching, preparation and follow-up lasts a total of around 60 min.

HAL therapy requires regular adjustment from the treating therapists to the patient’s current constitution. The many different variables in the setting (torque support, sensitivity of the sensors and balance between support in extension/flexion) of each individual joint of the HAL system result in many different setting options that can be selected depending on the form of the day or the objective of the respective training session. Adjustments can be made from day to day as well as within a training session. The basic aim is to gradually reduce the support that the system offers the patient.

HAL therapy can be further adapted by varying the use of the treadmill (speed, incline and decline) or by incorporating special exercises such as overcoming obstacles, stand-up exercises or walking backwards.

The control intervention consists of therapies that are eligible for prescription according to therapeutic products directive. For SCI, this includes the following options: physiotherapy, physiotherapy in a therapy pool and central nervous system physiotherapy (Vojta, Bobath and proprioceptive neuromuscular facilitation (PNF)), as well as the following ergotherapies: sensorimotor-perceptual treatment, motor-functional treatment and psychological-functional treatment.

The Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften (AWMF) guideline ‘Rehabilitation of the Lower Extremity, Standing and Walking Function in People with Spinal Cord Injury’²⁶ defines the following goals for physiotherapy:

1. Promotion of strength and endurance.

2. Tone regulation.

3. Improvement of mobility.

4. Coordination and motor learning.

However, there is no therapy with a clear recommendation or level I evidence. In established physiotherapy during the chronic phase following SCI, patients are typically treated with various therapies from the treatment catalogue. The goal is to preserve physical condition for as long as possible as a preventive measure and to prevent contractures or increased spastic muscle tone. Functional improvement can be achieved through standing and walking exercises using various devices (parallel bars, forearm walkers and rollators) and/or through the use of orthoses (foot drop orthoses and knee bridging orthoses).

The HAL therapy is primarily associated with aforementioned area (4), and additionally with areas (1) and (3). To determine the specific effect of HAL therapy in the study, the majority of the control intervention (>50%) will also be chosen from area (4), with relevant portions from areas (1) and (3). Since, as previously mentioned, there is currently no evidence-based standard therapy, the specific selection will be left to the therapists, taking into account the individual conditions of patients. Prior to the study, training will be provided for the participating therapists regarding the categorisation of possible therapies into the defined areas.

To assess the specific effect of HAL therapy, the control intervention is designed in analogy to the HAL intervention. A therapy algorithm consisting of five training sessions per week (Monday to Friday) over a period of 12 weeks, totalling 60 training sessions, is planned.

Similar to HAL therapy, each training session will last approximately 60 min, including stretching and preparation and follow-up activities. Documentation of the therapies performed will be completed weekly.

The control intervention will be conducted at the same rehabilitation centres as the HAL intervention.

If, for medical reasons (eg, acute illness), one or more therapy sessions cannot take place as scheduled, it is generally possible to reschedule them immediately after the planned end of the intervention. A maximum of 10 sessions may be rescheduled within a 3-week period. For subsequent analysis, various analysis sets are defined: the Full Analysis Set (based on the intention-to-treat strategy) consists of all randomised patients according to group assignment, and the Per Protocol Set includes all patients from the Full Analysis Set without serious protocol violations. The following protocol violations are classified as serious: discontinuation of the assigned intervention and use of the respective therapy for less than 45 sessions.

Regardless of study enrolment, patients are to continue their ongoing therapies according to the treatment guidelines, particularly for the upper extremities and the hands in patients with cervical SCI, as usual. This may include occupational therapy, physiotherapy and, especially, neurophysiological therapies such as Vojta, Bobath and PNF.

Medication for the treatment of spinal spasticity should also continue if it was already in place at the beginning of the study. Any changes during the study must be documented.

Due to randomisation, it is expected that supportive and regenerative treatments such as massage, stretching and detonisation for the lower extremities will be comparable in frequency and duration at baseline across both study arms. Regular monitoring of the data will be implemented to ensure this remains consistent throughout the study duration.

Rehabilitative treatments or training for the lower extremities according to the treatment guidelines will be suspended during the intervention to avoid overexertion.

Outcomes

The primary endpoint is the improvement in walking ability at the End of Treatment (12 weeks after randomisation) in the form of a combined endpoint. Therapy response is defined as at least one of the following three criteria being met:

1. Improvement in WISCI II score by ≥2 points.

2. Improvement in speed in the 6MWT by ≥0.1 m/s.

3. Improvement of the speed in the 10MWT by ≥0.15 m/s.

Since both endurance performance (6MWT) and short-term walking speed (10MWT) are important for the patients, it was decided to define the response based on these three criteria. By including the WISCI II score,²⁷ an improvement with regard to the required walking aids is taken into account. In case of a change of the necessary walking aid compared with baseline, the 10MWT will be performed twice with the baseline and the current walking aid, respectively. However, the fastest walking time will be included in the analyses. For patients who are unable to walk, the speed is set to 0 m/s. The response criteria for the 6MWT and the 10MWT are based on the clinically relevant improvements proposed by Forrest et al,²⁸ and for the WISCI II score on the clinically relevant improvements proposed by Scivoletto et al.²⁷

Secondary endpoints are the treatment response 3 months and 6 months after End of Treatment, the improvement in walking distance in the 6MWT and walking time in the 10MWT, as well as the improvement in the WISCI-II score. In addition, the AIS according to the ISNCSCI classification²⁹ will be determined at End of Treatment and 6 months after End of Treatment, as this was considered a core outcome according to both the AWMF S2e guideline, ‘Outcome assessment in initial treatment after newly acquired spinal cord injury’ and the Spinal Cord Injury Evaluation Database to Guide Effectiveness Task Force.^{30 31} In addition, the change in spasticity is examined at all visit times via the modified Ashworth Scale.³²

Furthermore, the following patient-reported outcomes will be collected: Quality of life (WHOQOL-Bref³³), pain and neuropathic pain (Numeric Rating Scale/Spinal Cord Injury Pain Instrument³⁴/Neuropathic Pain Scale NPS³⁵), bladder function (Short Form-Qualiveen³⁶), rectal function (Neurogenic Bowel Dysfunction Score (NBD-Score)^37–39) and concomitant therapies/medication.

The selection of instruments was guided by the recommendations for core outcomes from the AWMF S2e guideline ‘Outcome Assessment in initial treatment after newly acquired spinal cord injury’, along with the associated systematic review and international recommendations for core outcomes.^{30 31 40}

Harms

The study compares conventional therapy according to the therapeutic products directive with neuromuscular feedback therapy using the HAL system. The HAL system is a medical device with CE marking (CE 0197) and is approved in the European Union for patients with SCI, traumatic brain injuries, cerebrovascular diseases and disorders of the brain and neuromuscular system. To date, no severe adverse events (AEs) have been observed with the use of the HAL system. However, potential transient adverse effects may arise from its use, including:

• Pressure sores in the areas of the ankles, knees and hips. Additionally, poorly fitted footwear may lead to pressure injuries on the feet.

• There is also a theoretical risk of muscle and tendon ruptures, as well as fractures, due to the force of the motors.

• Patients are connected to the device via cables and electrodes, which are powered by a battery, posing a risk of electrical shocks.

Possible AEs that may occur in both intervention groups include:

• Falls during training. To prevent this, patients are secured with a waist belt and suspension system, but ultimate safety cannot be guaranteed. There is a particular risk of falls during the transfer from the wheelchair to the treadmill. In the event of a fall due to loss of strength, there is a risk of bruises, lacerations and fractures.

• Due to cardiovascular strain during training, patients may experience chest tightness, and there is a theoretical risk of a heart attack. Therefore, preventive monitoring of pulse rate and blood pressure will be conducted throughout the training.

• The intensity of the training may lead to exhaustion and, in particular, muscle soreness.

All occurrences defined as AEs as well as severe AEs (SAEs) will be recorded weekly in the study database and evaluated regularly.

During the study visits, there will be no fall protection during the functional walking tests; patients will use their usual assistive devices, thus presenting a risk of falls with the aforementioned injury risks. However, these tests are standard assessments both in clinical practice and clinical trials, so this risk is considered negligible.

According to the ethics committee’s requirements, liability insurance for study participants will be obtained. To monitor safety-related events, a data safety and monitoring board (DSMB) will be established, which will evaluate the safety-related events of the study once a year and provide recommendations for any necessary adjustments to the study management.

Participant timeline

The participant timeline with information on data collection is presented in the SPIRIT figure (table 1).

Table 1. SPIRIT figure participant timeline.

	V0	Rand	V1	V2	V3	V4
Time point (w/m)	−1 w	0	+6 w	+12 w	+6 m	+9 m
Study site (SCI/RC)	SCI	RC	RC	SCI	SCI	SCI
Patient enrolment and group allocation
Screening according to inclusion/exclusion criteria	x
Informed consent	x
Randomisation		x
Intervention and follow-up
Study intervention (neuromuscular feedback therapy)
Control intervention (conventional therapy)
Assessments
Sociodemography	x
Primary endpoints
10 m walk test	x		x	x	x	x
6 min walk test	x		x	x	x	x
WISCI II score	x		x	x	x	x
Secondary endpoints
AIS according to ISNCSCI	x			x		x
Spasticity (modified Ashworth Scale)	x				x	x
Patient-reported endpoints
Quality of life (WHOQOL-Bref)	x		x	x	x	x
Bladder and rectal function (SF-Qualiveen/NBD score)	x		x	x	x	x
Pain and neuropathic pain (NRS/SCIPI)	x		x	x	x	x
Concomitant therapies/medication	x				x	x
Safety
AEs, SAEs					x	x

AE, adverse event; AIS, American Spinal Injury Association Impairment Scale; ISNCSCI, International Standards for Neurological Classification of SCI; m, months; NBD, neurogenic bowel dysfunction; NRS, Numeric Rating Scale; RC, rehabilitation centre; SAE, severe adverse event; SCI, spinal cord injury; SCIPI, Spinal Cord Injury Pain Instrument; SF, Short Form; SPIRIT, Standard Protocol Items: Recommendations for Interventional Trials; w, weeks; WHOQOL, WHO quality of life; WISCI, Walking Index for Spinal Cord Injury.

Statistical methods

The aim of the study is to compare the two therapies in terms of their application as a treatment strategy. Therefore, the comparison of the response rates will be conducted without considering potential therapy discontinuations or compliance issues. For this reason, the analyses will be performed in the intention-to-treat population. The following pair of hypotheses will therefore be tested for the primary endpoint using a two-sided significance test at the 5% level:

$${\displaystyle {\displaystyle {\mathrm{H}}_0:\mathrm{p}\mathrm{H}\mathrm{A}\mathrm{L}=\mathrm{p}\mathrm{P}\mathrm{T}\mathrm{v}\mathrm{s}.{\mathrm{H}}_1:\mathrm{p}\mathrm{H}\mathrm{A}\mathrm{L}\ne \mathrm{p}\mathrm{P}\mathrm{T}}}$$

Here, pHAL and pPT represent the response rates in the feedback therapy group and the control group, respectively. Secondary endpoint analyses will also be conducted using two-sided tests at a significance level of 5%. The statistical analyses are carried out using the statistical software R (www.r-project.org, R Foundation for Statistical Computing, Vienna, Austria), V.4.1.0 or higher.

Planned analyses

The analysis of the primary endpoint will use a generalised linear mixed model with a logit link (mixed logistic model). The model will include fixed effects for the intervention group, as well as for the 6MWT and 10MWT speeds at baseline, the WISCI II score, age, time since onset of the SCI and spasticity as covariates. The effect of the rehabilitation centre will be included as a random effect in the model. The parameter of interest is the OR for the treatment response between the HAL and control groups. Estimation will be performed using the restricted maximum likelihood method, along with the provision of a 95% CI.

The exploratory analysis of secondary quantitative endpoints will be conducted using analogous generalised linear mixed models with a linear or logit link depending on the measurement level (linear for all quantitative endpoints, such as quality of life, walking distance in the 6MWT, speed in the 10MWT and logit for dichotomous endpoints). For quantitative endpoints, the baseline outcome values will also be included as covariates in the model unless they are already present. In the case of repeated measurements of the outcome at various time points, such as for quality of life, the analysis will use a model that accounts for these repeated measures. An additional random patient effect and a fixed interaction effect for therapy group×time will be included in the models. The intervention effect at individual time points will be determined using the least square means method. All analyses will be conducted blinded to group allocations.

Missing values should be replaced by multiple imputation when analysing the primary endpoint. In addition, sensitivity analyses with the most conservative possible replacement of missing values should be performed for the primary endpoint. Details of planned analysis will be written down in the statistical analysis plan before database look. No interim analysis is planned.

Data monitoring committee

This is a multicentre study. Safety is monitored centrally. There is a DSMB that meets once a year and evaluates all AEs and SAEs reports. In the event of an accumulation of study-related AEs, the DSMB will be convened at an early stage. The DSMB is composed of individuals with relevant expertise, including a biostatistician, an orthopaedic surgeon and a trauma surgeon. The members of the DSMB are independent of the conduct of the study and free of conflicts of interest.

Patients can withdraw from the study at any time without giving reasons and without suffering any disadvantages. The study physician can also initiate withdrawal in the event of medical risks. However, withdrawal from the intervention (eg, due to health problems or lack of participation) does not necessarily lead to withdrawal from the study – further participation in the study is possible. The entire study may be terminated if, for example, too few participants are recruited, serious side effects occur or ethical concerns arise. Individual study centres may be excluded if, for example, they provide insufficient documentation or violate guidelines.

Trial monitoring

Alcedis GmbH (Gießen, Germany) as the contracted research organisation involved carries out monitoring. It is the authorised representative of the sponsor and is responsible for monitoring the proper conduct of the clinical trial and quality control at the trial site. The results are recorded electronically in corresponding monitoring reports within the electronic Case Report Forms.

Data management

The institutions conducting the study will comply with all relevant data protection regulations and ensure that the necessary technical and organisational measures are implemented. Research data will be pseudonymised at the earliest possible stage (as soon as it is entered into the study platform by the study centres). Within the scope of the study, the personal data of patients will only be processed in compliance with the relevant data protection regulations, that is, only to the extent necessary for the conduct of the study. The processing of the personal data of patients will only take place after they have given their verbal and written consent to the collection, transmission and evaluation of their data. The data is protected against unauthorised access. The contact information of study participants is stored securely in the SCI centres and rehabilitation centres for internal use (appointment scheduling/rescheduling, inquiries) during the study. The data processed by the Alcedis-Electronic Data Capture (EDC) systems is stored in a relational database with restricted access. The data and EDC systems are protected by mechanisms such as intrusion prevention. Encrypted backups are created daily and weekly. The daily backups are mirrored in a second data centre to prevent data loss in the event of a local incident. The weekly backups are stored in a fireproof safe until they are delivered to a special tape data archive. Alcedis GmbH is responsible for data processing in the project.

Discussion

Participants in both study groups are considered patients with chronic SCI who have completed treatment by the time they enter the study. At this stage, therapies to control symptoms can still be prescribed. Although intensive therapies involving daily training sessions could still be prescribed in accordance with the therapeutic products directive, they are usually no longer carried out as functional improvement is no longer expected. In this respect, it is assumed that patients in both groups will benefit from participating in the study, although the benefits are expected to be superior in the HAL intervention group. Improvements in uprightness and joint mobility, increased muscle development, and a possible training-induced improvement in cardiovascular activity are hypothesised. In the HAL intervention group, functional and neurological improvements can be expected, as well as an improvement in walking ability.

If this study can prove the superiority of therapy with the HAL exoskeleton, this therapy could become available as a health insurance benefit for all patients with fully treated chronic SCI and preserved residual motor function. This would have a positive effect on these walking ability and quality of life of patients.

Protocol version

This paper is based on Protocol V.1.3, 6 March 2026.