Exoskeleton Training for Spinal Cord Injury Neuropathic Pain (ExSCIP)

Background and Rationale:

Following SCI, approximately 53% of people develop neuropathic pain (NP). Irish SCI data identifies high pain intensity and pain interference levels with NP and significantly poorer quality of life (QoL) than other pain phenotypes. Individuals can describe NP as more debilitating than the other consequences of SCI, as their most persistent health issue and adequate pain relief as an unmet need.

International data identify the proportional burden of NP following SCI as significant. Ninety-four percent of individuals are prescribed >1 medication, the mean number of physician office visits in a 6-month period due to SCI NP is reported as 2 and the total annualised cost of NP per subject in the United States (US) is reported as $26,270 (direct $8,636, indirect $17,634).

The presence of pain is further associated with lower return to work rates following injury, and more than a third of individuals with SCI in employment report frequent pain interference with their work . Pain interference with function, health status and work are noted to be significantly worse in individuals with more severe NP, where overall work impairment is reported at 38%.

NP after SCI is multi-faceted and heterogenous, making isolation of specific mechanisms more challenging. Mechanisms hypothesised for NP after SCI include neuronal hyperexcitability (central and peripheral sensitisation) and corticothalamic maladaptive neuroplasticity. Additionally, NP symptom severity post SCI has been reported to be associated with a combination of residual spinothalamic tract (STT) function below the level of injury and with catastrophising pain coping mechanisms.

The mechanistic effects of sensorimotor stimulation on NP stem from Phantom Limb Pain research (PLP) with significant reversal of cortical dysfunction in the primary somatosensory cortex of individuals with PLP evident. Similar maladaptive cortical reorganisation is hypothesised to be associated with NP in SCI.

This is further supported by data garnered from electroencephalography (EEG) studies showing that changes in oscillatory brain activity known as thalamo-cortical dysrhythmia, are associated with the presence of NP. NP in SCI is associated with an EEG power signal increase in the theta band and possibly high beta band but a decrease in the high-alpha-low-beta band. In addition, NP in SCI is associated with decreased reactivity of alpha band power signals in response to eye opening. Thus it has applications as a biomarker for current NP and as a predictor of development of future NP.

The mainstay of NP treatment after SCI is pharmacotherapy with anticonvulsants and antidepressants to reduce pain intensity. Pregabalin/gabapentin, duloxetine, amitriptyline and/or opioids are the first- and second-line treatments recommended, although severe pain remains refractory to these treatments in 2⁄3 of sufferers. Survey data report high use of non-steroidal anti-inflammatories and paracetamol.

Significant side-effects of medications are reported. SCI patients are particularly prone to central nervous system related side effects which are often intolerable. These, together with fear of medication dependency, result in poor adherence to pharmacological regimens leading to a call for non-pharmacological treatment options for people with NP after SCI.

Virtual reality (immersive virtual walking virtual illusion/imagined walking) has shown promise for reducing NP intensity and interference after SCI. Virtual illusion interventions show evidence of direct and corrective stimulation to the reorganised sensorimotor areas in SCI patients with NP, supporting the theory that NP mechanisms are reversible. However, actual sensorimotor intervention studies are inconclusive in SCI at this point with limited focus on walking despite compelling preclinical studies showing prevention and/or reversal of SCI neuropathic pain. Notably in animal studies, other exercise paradigms including swimming and stance training had only transient or no effects on SCI-induced NP suggesting that the rhythmic stimulation of proprioceptive and mechanosensory afferents together with weight bearing experienced in walking might be necessary to reduce NP.

The exoskeleton intervention itself is not new within the neurorehabilitation space for SCI. However, no RCTs to date have specifically recruited participants with moderate-to-severe NP in order to assess its mechanistic effects on NP. The ExSCIP randomised feasibility trial addresses this current knowledge gap, examining exoskeleton-based walking 3 times per week, as a mechanistic-based intervention for NP after SCI. It will test the feasibility and acceptability of an exoskeleton, and whether it demonstrates positive signals in reduction of NP intensity and interference levels to warrant onward progression to a definitive trial.

Aims and Objectives:

The overall aim of this study is to examine the feasibility and acceptability of an exoskeleton, mechanistic-targeted, walking intervention for NP in people with SCI.

The primary objectives for the study are:

1. Implement an exoskeleton training programme for people with below level NP > 6 months after a traumatic SCI.
2. Pilot and assess the impact of an exoskeleton-based walking intervention in NP > 6 months after SCI, examining feasibility outcomes and short and long-term (6 months) changes in pain intensity and pain interference.

The ExSCIP study is a phase 2 randomised, single blinded, feasibility trial with the aim of examining progression criteria for a definitive trial.

Progression criteria are based on consideration of the primary objectives around feasibility and the potential for effectiveness and implementation in clinical practice. Quantitative and qualitative process evaluation data will be analysed to consider the following continuation criteria.

• Successful uptake, recruitment, and retention.
• Successful implementation of the ExSCIP intervention.
• Process evaluation indicates that ExSCIP is acceptable to people with NP after SCI and to staff delivering the intervention.
• A positive effect on pain and pain interference outcomes are identified and are meaningful.
• Cost analysis indicates that the ExSCIP intervention might be cost effective. The intervention will be delivered in the Motion Analysis Laboratory at University College Dublin (UCD).

Participant Screening:

Stage 1: Phone Screening:

NP will be screened for as a minimum criterion initially by phone. This phone screening will do the following:

• Confirm their SCI diagnosis (e.g. traumatic aetiology and >6 months post injury).
• Confirm they are on a stable medication regimen.
• Confirm they are exoskeleton naïve.
• Screen for the presence of NP using the Spinal Cord Injury Pain Instrument (SCIPI).
• Inquire into anthropometric details, e.g. the candidate's height and weight to give further indication relating to their potential suitability/compatibility with the exoskeleton.
• Once the phone screening has been completed, candidates deemed to be potentially suitable to participate in the trial will be booked in for an in-person assessment. Potential participants will be provided with a study information leaflet at this point and informed consent will be sought from participants to complete an in-person assessment. A 1 week grace period will be given to participants between provision of the study information leaflet and obtaining informed consent.

Stage 2: In-Person Assessment:

An in-person assessment to confirm participant suitability will be performed by an independent assessor. The assessment will entail the following steps:

Confirmation of presence of moderate to severe below level NP:

• NP will be confirmed based on a neurological examination, a score of ≥4 on the Douleur Neuropathique 4 (DN4) and a comprehensive pain history.
• This will be supported by the use of the ISCIP Pain Classification.
• Moderate and severe NP as confirmed above will be described as pain ≥ 3 and ≥ 6 on the 0-10 Numerical Rating Scale (NRS) for NP (averaged over a week).

Anthropometric and clinical assessment for compatibility for use of exoskeleton:

• Participants will undergo an anthropometric assessment to ensure no height, weight, joint range of movement or muscle spasticity restrictions to exoskeleton use apply.

Stage 3: Informed consent and data collection:

• Candidates who meet inclusion criteria will be provided with a study information leaflet (see Figure 2) and asked to provide written informed consent that they agree to participate in the study via a consent form.
• The independent assessor will then collect data at baseline for the outcome measures outlined: (Please refer to outcomes section for full details):

Data/statistical analysis:

Descriptive statistics and estimation using 95% CIs will be the main focus of the analysis. The number of participants recruited and retained, and information on missing or incomplete data from all outcome measures will be explored. Baseline demographics and outcome variables will be compared at all assessment times within groups.

For categorical measures, frequencies and percentages will be presented and for continuous measures, the mean and standard deviation (SD) will be reported. For continuous measures which show evidence of some skew a median and interquartile range may also be presented or substituted for the mean and SD. Within group change scores and their 95% CI will be examined in relation to the MCID. Repeated measures ANOVAs will be used to compare between group differences of continuous variables across the three time points. Statistical significance will be determined α-priori at an alpha level of 0.05.

For analysis of EEG data, this will be an exploratory analysis using a multilevel linear mixed model (LMM) approach to examine differences between the intervention groups over time in the EEG alpha, beta and theta band power. Repeated measures within participants will be modelled as a random effect. Fixed effects in the model, will include group assignment and time. The moderating effects of pain intensity and interference will also be evaluated. The LMM will study both main effects and interaction effects using the R package lme4 to fit the models. Models will be compared using Likelihood Ratio Tests (LRT) to assess the significance of effects. Statistical significance will be determined α-priori at an alpha level of 0.05.

When all data is collected, data analysis will be conducted by a data processor blinded to group allocation. A full statistical analysis plan will be prepared prior to final analysis. Statistical analysis will be conducted using SPSS version 29 software and analysis will be conducted as intention to treat (ITT) and per protocol.