Background: In Switzerland, about 6000 individuals live with the consequences of a spinal cord injury (Brinkhof et al, 2016). One of the major goals after an incomplete spinal cord injury (iSCI) is to regain walking function. To this end, different approaches are used in rehabilitation such as treadmill-based, robotic-assisted (exoskeleton or end-effector) and conventional gait training. According to current literature, the superiority of one of these approaches remains unclear (Mehrholz, Harvey, Thomas, and Elsner, 2017); In the research on gait rehabilitation after iSCI, recent randomized clinical trials (RCTs) found no statistical differences between conventional gait training and robotic-assisted gait training. Nevertheless, according to the comparison of effect sizes obtained from these training, these trials suggested that the conventional training approach leads to larger improvements in gait capacity when compared to robotic-assisted therapy (Field-Fote and Roach, 2011; Nooijen, Ter Hoeve, and Field-Fote, 2009). Therefore, these trials highly recommended further research considering these aspects. However, in clinical settings, the implementation of such systematic and intense training sessions remains challenging. The present study aims to test the hypothesis that conventional training might have larger effect sizes on gait capacity and to evaluate the feasibility of such systematic training in a clinical setting of inpatient rehabilitation.

Objectives: To contribute to the current knowledge on best clinical practice in gait rehabilitation within the iSCI population. More specifically, the study objectives are two-fold: A first objective is to compare the effects of conventional training, end-effector based therapy and the combination of these interventions on the gait ability of iSCI. A second objective is the evaluation of the feasibility of systematic gait training protocols in a clinical setting.

Participants: Individuals with motor incomplete spinal cord injury (iSCI), presenting a traumatic or non-traumatic iSCI with an injury onset <6 months.

Intervention: Participants will be trained in one of the three groups by trained physical therapists during 10 sessions, 3x/week with an average duration of 30 minutes.

Outcomes: To attain the first objective the effects will be quantified by the following main outcomes: Walking capacity (independence), walking speed, and safety. Feasibility of the systematic intervention will be evaluated using the drop-outs of therapy interventions.

Background and Rationale:

The present study aims to add evidence for best practice in gait recovery interventions for individuals with incomplete SCI (iSCI). In contrast to post-stroke rehabilitation, the evidence on best practice for gait recovery for individuals with iSCI is less well-founded. A recent Cochrane review and the guidelines from the Deutsche Gesellschaft für Neurorehabilitation suggests that in post-stroke rehabilitation the combination of conventional gait training (walking overground) and robotic-assisted gait training is most successful to improve gait capacity (Mehrholz et al., 2014; Dohle et al., S2e Leitlinie ReMoS, 2015); Furthermore, guidelines suggest according to the high level of evidence, that the use of robotic-assisted training is advantageous to improve walking speed and endurance in individuals with post-stroke.

Such recent guidelines for spinal cord injury do not yet exist to knowledge of the investigators. In 2014 a Cochrane Review was published indicating that there is insufficient evidence about the best practice for gait recovery in incomplete SCI individuals (Mehrholz, Pohl, Elsener, 2014). Recently, a systematic review points out the critical assumption that according to the present evidence robotic-assisted gait training is not superior to conventional gait training (walking overground) (Mehrholz et al, 2017).

Since these studies suggest a high frequency of training and systematic intervention over several weeks (ex.: 4-5x / week; over 3 weeks) the present investigators would like to conduct a randomized-controlled trial with a two-fold aim: 1. Adding evidence to existing knowledge on gait rehabilitation best practice in individuals with iSCI. And 2) assessing the feasibility of such a protocol in a clinical setting in daily clinical practice. The latter is a very important aspect considering that the transfer from research knowledge into clinical practice is a large challenge (Maher et al, 2004). Among the most frequent barriers to implement evidence into practice are methodological problems (66%), a lack of replication (54%) and poor generalisability (53%) according to the analysis by Metcalf and colleagues on 715 clinicians (physio-, occupational, speech and language therapists) (Metcalfe et al, 2001).

Study design:

randomized, pre-post design, pilot-study to test the superiority of conventional gait training to mechanical-assisted gait training in individuals with motor incomplete spinal cord injury at a sub-acute stage.

Outcomes for patients' characteristics:

Additionally, to the outcomes measures of gait ability participants will be assessed for clinical parameters on structural and activity level to characterize participants level of impairment and deficits. The assessments in question quantify lower limb spasticity and motor function.

Spasticity will be assessed with the composite spasticity index (CSI) will be used to test spasticity in the lower extremities. The CSI is a 16-point scale measuring the reflex response for example after a tap on the Achilles or patellar tendon (4 points), resistance to full-range passive ankle dorsiflexion at a moderate speed (8 points), and duration of the clonus at the ankle (4 points). Value intervals of 1-5, 6-9, 10-12 and 13-16 correspond to absent, mild, moderate and severe spasticity, respectively (Levin and Hui-Chan, 1993).

The motor function will be quantified using the International Standards for Neurological Classification of Spinal Cord Injury, revised 2011, published by the American Spinal Injury Association (ASIA) (Kirshblum and Waring, 2014).

Safety Outcomes: For safety purpose, a 16-point Borg Scale will be assessed immediately prior, during (2-3x) and immediately after the training session. The cut-off for training cessation is set at a Borg score of 15-16 (hard) (Swain et al, ACSM 2014). According to literature in healthy a score of 15-16 indicates that the anaerobic threshold has been exceeded. Since the training should in a submaximal level the aim is to train participants in an aerobic range of energy supply.

Furthermore, potential safety outcomes are muscle soreness or pain induced by skin irritation of the harness used with the end-effector device. Skin irritation will be inspected and controlled by the nurses responsible for participants as it is conducted regularly and systematically as an in-house practice.

Methods

Recruitment:

Participants will be recruited at the local inpatient rehabilitation of the REHAB Basel, Switzerland. Once participant accepted participation, they will be randomly assigned to one of the three groups with the use of the randomization process for block randomization on www.randomization.com (Dallal E.G.; last access 6.05.19). The allocation is conducted from externally. Assessors are blinded during the process. Participants and Physiotherapists cannot be blinded due to the training location.

Screening:

At admission to the REHAB Basel in the ward for individuals with spinal cord injury, each patient was screened by the assessors from the rehabilitation centre. Medical Doctors responsible for the ward are informed about the in- and exclusion criteria and can report a potential participant for screening. Screening includes the consultation of the medical history (digital) checking for in- and exclusion criteria particularly lesion location and onset. If potential participants pass the screening of medical dossier the assessors communicate with the responsible therapist about the inclusion criteria on the capacity to perform sit-to-stand and step initiation.

The pre-selected participants are then invited for a first visit to give the patient information and explain and hand-out the informed consent. In case of the incapacity to sign, secondary to the lesion, participants are asked to a legal relative (mature) signing the paper.

Once the informed consent is signed the participants were planned for a pre-evaluation to test the final eligibility with the gait assessments and clinical evaluation. In case the participant does not meet inclusion criteria or stop the participation there will be no consequences on participants' treatment. Study participants do not receive any compensation or payment.

Assignment to study groups:

Randomization is conducted by the principal investigator. The investigator is not involved in the pre- and post-evaluation of the participant and not on site. Before the study, the subject numbers have been assigned to the groups using the online calculating system for block randomization www.randomization.com. With the subject's number assessor obtain the information about the group assignment via E-mail.

Concealing of group assignment: The group assignment is registered in an xls file where the ID of the participants is encoded and some demographics are registered. The folder with the xls file is registered in a personnel file of the assessors called RCGTSCI_18 (robotic-conventional gait training spinal cord injury 2018).

Blinding:

The physiotherapists conducting the treatment cannot be blinded since they have to know all of the three treatment types since they can be assigned to every patient. The assignment of physiotherapist (PT) to patients is based on the availability in the PT team.

Participants have to be informed a priori about all the potential treatments they can be assigned to. Based on treatment nature participants cannot be blinded.

The assessors are blinded. They are not involved in the training of participants or the prior randomization.

Criteria for withdrawal: Should a participant be expelled prematurely due to a particular reason for the cancellation criteria or by the withdrawal of their consent before the end of their participation, this participant will be replaced by a new participant. The data of the expelled participant will not be used for any analysis.

Study procedures:

As mentioned above, the patients are personally informed about the study. Afterwards, they are neutrally asked about their willingness to participate. They will receive an information sheet and a declaration of consent. Patients are informed that they have time for consideration and that they will be contacted again in 2 days.

When patients give consent to the study, they are randomly assigned to a group by an uninvolved person. The central therapy planning is informed about the study participation of the person concerned and then plans the study procedure.

The first baseline measurement is planned 2 days before the start of the intervention. Main outcome parameters (FAC, CSI, MFW, TUG and 10mWT) are recorded during the measurement. After the measurement, the further procedure depends on group allocation.

Pre-evaluation Screening Visit (Day 0): The possible participants are visited by one of the authorized assessors. Participants will be informed verbally about the study project and receive the information sheet and the declaration of consent. The potential participants are informed that they have time to decide and that they will be visited again in the next 2 days for clinical evaluation.

After the participant gave the consent to the study, the participants get randomly assigned to one of the three groups; Appointment 1 Pre-evaluation (Day 2): The authorized examiners test the composite spasticity. They evaluate the motor score of the lower extremity, the functional ambulation categories, 10-meter walking test and the timed up and go test (described in 9.2.) Post- and Follow-up evaluation Appointment 12 (day 1-2 after the last training session): During post-evaluation the main outcome parameters for gait ability, gait speed and safety are assessed by the authorized assessors. The outcomes are quantified with the functional ambulation categories, 10-meter walking test and the timed up and go test (described in 9.2.) in the same manner as during pre-evaluation. Clinical evaluation of spasticity and motor function is not conducted since these parameters do not count as outcome parameters. These assessments serve to characterize the participants.

There is no follow-up survey planned. The records of participants who prematurely stop the study are not used for analysis.

All clinical evaluation measurements are carried out by 3 trained physiotherapists. The intervention units are carried out by certified, experienced physiotherapists. The therapists record the content and the distance as well as all values to be documented.

Statistical considerations:

Hypothesis:

This pilot study aims at first to test if conventional gait rehabilitation (CT) is superior to robotic-assisted therapy (EET) or the combination of both (CETcomb), conventional and robotic-assisted therapy as suggested by recent literature. In line with that, time-effect of training (pre- post-training) for all three interventions on gait ability (FAC), gait speed (10MWT) and safety (TUG) will be analysed.

In line with these objectives the hypotheses are formulated as follows:

H0: Group A = Group B = Group C; Group A (EET), Group B (CT), and Group C (CETcomb) will all result in significant improvements in gait ability, speed and safety.

H1: Group C (CETcomb) > Group B (CT) > Group A (EET); The combination group (Group C) will result in significantly higher improvements in gait ability, speed and safety from pre- to post-testing when compared to Group B and Group A. The time effect in Group B will be significantly higher when compared to Group A.

Determination of Sample Size:

For the present pilot study, a participant number of at least 12 was defined for each group based on a sample size calculation (Noordzij et al, 2010) using the mean and standard deviation of walking speed found in individuals with motor incomplete spinal cord injury. The number of 12 participants in each group was necessary to achieve a power of 0.80. The power was set at 0.80 which is considered as an optimal power to avoid false rejection of the H0 hypotheses.

For the formula 0.05-0.06 m/s was used as SD2 since this range was considered as minimal clinical relevant change (Field-Fote and Roach, 2011; Musselman, Fouad, Misiaszek, and Yang, 2009; Street and Singleton, 2018) Consequently, the target difference for the present study was set at 0.06 m/s.

Planned Analyses:

The planned analysis consists of a hypothesis testing for time-effect and group comparison in line with standard treatments in gait rehabilitation. The data set consist of clinical data obtained with standardized assessments in physiotherapy (i.e. gait assessments and evaluation of clinical parameters such as muscle strength and spasticity). The values obtained are on an ordinal, interval and ratio level. The analysis will be conducted after the completion of data collection of the aimed 36 participants (12 per group).

Statistical analysis:

For primary outcomes, time effect on gait parameters for each group will be analysed at two-time points: pre to post-training. Furthermore, the difference between pre- and post-evaluation at the end of data collection for each participant respectively. These values will be used for statistical comparison. These differences will be tested for the normality of data distribution. Therefore, the Shapiro-Wilk test will be applied to the difference between pre- and post-evaluation.

The level of significance is set at an alpha level of ≤0.05. The secondary outcome is the feasibility of the training protocols in a clinical setting. Feasibility will be quantified with the number of drop-outs and number of completed training sessions (total of 10 sessions per participants). The Study is considered feasible if 80% of the participants completed >8 sessions of training (80% of total sessions).

The analysis will be conducted by the investigator, Martina Betschart, using the current version of SPSS (23.).

Inclusion Criteria:

• Incomplete paraplegia or quadriplegia and classification of C and D in the AIS Score from the American Spinal Injury Association (ASIA) (Kirshblum and Waring, 2014).
• The participants should be able to get from a sitting position to a standing position with the aid of an assistant and bars
• Participants need to have a stable cardiovascular condition, absence of known heart disease or a known heart disease classified as class I or II in the New York Heart Association Classification system (American College of Sports Medicine; Guidelines for exercise testing, 2014); medical history is verified by the medical doctor from the SCI ward the REHAB Basel.
• Sufficient compliance for the treatment action
• And have signed the informed consent for the present study

Exclusion Criteria:

The presence of any one of the following exclusion criteria will lead to exclusion of the participant

• Patients with an ASIA classification of A and B
• Patients who are unable to stand up from a chair with moderate or no personal assistance
• Patients diagnosed a high-grade osteoporosis
• Patients with severe lower limb movement restrictions (contractions or fractures)
• Known or suspected non-compliance, drug or alcohol abuse,
• Inability to follow the procedures of the study, e.g. due to language problems, psychological disorders, dementia, etc. of the participant

The individual data collected will not be shared with other researchers.

The assessors are blinded since the testing assessor is not conducting the randomization. Patients and the physiotherapists conducting the training are not blinded. This is not due to the training location.