Edited by: Ronaldo M. Ichiyama, University of Leeds, United Kingdom
Reviewed by: Karen Minassian, Medical University of Vienna, Austria; Jamie Lynn Reed, Vanderbilt University Institute of Imaging Science, United States
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Almost 800,000 people suffer from traumatic spinal cord injury (SCI) worldwide every year (
Epidural spinal cord stimulation (eSCS, SCS, or estim) has long been used for the treatment of chronic pain (
The mechanisms by which electrical stimulation restores supraspinal control or modulates the function of the spinal cord remain unclear. Careful electrophysiology during motor-control tasks has revealed subtle supraspinal control in more than 80% of participants with clinically motor-complete injuries (
While the potential biological effects of
The Epidural Stimulation After Neurologic Damage (ESTAND) trial tests the effect of eSCS after motor-complete thoracic SCI on volitional movement and autonomic function without implementing locomotor therapy (
All the procedures described in this study were approved by the Hennepin Healthcare Research Institute Institutional Review Board and with an Investigational Device Exemption from the United States Food and Drug Administration. Patients with chronic, traumatic SCI (more than 1 year since injury) were recruited if they met the following criteria: older than 22 years of age, ASIA Impairment Scale (AIS) classification A or B with a neurological level of injury between C6 and T10, full arm and hand strength and intact segmental reflexes below the level of injury. Participants were excluded if they had medical or psychological comorbidities that would significantly increase the risk of surgery, severe dysautonomia (systolic blood pressure fluctuation below 50 or above 200 mmHg) during autonomic testing, contractures, pressure ulcers, recurrent urinary tract infection, unhealed spinal fracture, recent botulinum toxin use, or pregnancy. Once enrolled, subjects were asked to suspend any medications used for spasticity, for example, baclofen and oxybutynin. This analysis includes seven participants that have completed 80% or more of the study. The six participants that have completed the study in its entirety were enrolled for a range of 1.26–1.47 years. Overall, participants had a mean age (±SD) of 42 ± 11.4 years, and a mean time since injury (±SD) of 7.7 ± 4.8 years ranging from 3 to 17 years (
Demographic information: *ages are rounded to the closest decade.
Subject | SVM | Age* | Years post injury | AIS | Lowest injury site | Vertebral fracture level | Mechanism of injury | Implant level |
1 | No | 50 | 10.96 | A | T8 | T8 | Fall | T11 |
2 | Yes | 30 | 8.21 | A | T4 | T4 &T5 | Sports injury | T12 |
3 | Yes | 40 | 16.83 | A | T8 | T7 &T8 | MVA | T12-L1 |
4 | No | 40 | 5.36 | B | T5 | T5 &T6 | Fall | T11-T12 |
5 | Yes | 50 | 5.44 | A | T5 | T5 | MVA | L1 |
6 | Yes | 60 | 4.02 | A | T5 | T4-T5 dislocation fracture | MVA | L1-L2 |
7 | No | 30 | 3.05 | A | T4 | T4 | Sports injury | T12-L1 |
All participants provided thoracic spinal cord magnetic resonance imaging (MRI) at screening (
T2 sagittal thoracic magnetic resonance imaging (MRI) obtained prior to enrollment.
All participants were implanted with an epidural stimulator upon enrollment after baseline data, including a detailed neurological exam and self-reported questionnaires, were collected. Participants underwent epidural placement of a three-column, 16-contact paddle lead through a T11–T12 laminectomy and subcutaneous placement of a primary cell internal pulse generator (IPG) (Tripole and Proclaim Elite, Abbott, Plano, TX, United States) in the lower lumbar area under general anesthesia (
Placement of 5-6-5 epidural paddle lead through a T12 laminectomy, overlying the T12–L1 epidural space. Final lead placement is guided by intraoperative electromyogram (EMG).
The BMCA was conducted during subject screening and twice at each follow-up visit, without and with eSCS. The eSCS program was selected based on participants’ preferences during the previous month of eSCS use, as well as objective data on the current month’s settings (
Electromyogram was pre-processed by removing 60-Hz noise with a Fourier filter and then power calculated in windows of time by average root mean square (RMS) (
The Muvi 300 cycle from MOTOmed was used to assess functional movement capabilities. The Muvi 300 includes a motor-assisted setting that facilitates training with minimal muscle strength by switching from passive to active training without strain. The active assist motor can vary in speed (rpm) and resistance. When the force sensitivity threshold is met, the motor ceases and the patients pedal on their own in the bike’s active mode. The bicycle collects trial duration [seconds (s)], active and passive mode duration (s), active and passive distance traveled [meters (m)], active and passive speed [rotations per minute (rpm)], work done in active mode [kilojoules (kj)], and average and maximum energy produced during active phase [watts (W)].
The factorial design bicycling task was added after initial participants demonstrated enough movement capabilities with active eSCS to allow for more robust functional assessments during follow-up sessions. Therefore, the data capture window for each participant differs based on the relative enrollment date. If the bike was implemented after a subject initiated eSCS therapy, they did not undergo a baseline visit.
At the baseline visit, participants completed two motor-assisted bike trials: one passive trial where the subject was asked to relax and one active where the subject was asked to attempt to pedal. At each subsequent visit, the subject completed the same passive and active trials without stimulation and additional passive and active trials with selected preferred stimulation setting on. The duration of each trial was 2 min. As a result, there were two conditions without stimulation analyzed: (1) no effort, no stimulation and (2) maximal effort, no stimulation.
Four participants unexpectedly developed sustained volitional movement with stimulation turned off, referred to as the SVM group. In order to ascertain group differences, participants from the SVM group were compared to the participants who only demonstrated movement with stimulation, referred to as the non-SVM group (
Differences in baseline Modified Ashworth Scale scores between the spontaneous volitional movement (SVM) group (Yes) and non-SVM group (No). Participants in the SVM group had significantly higher spasticity MAS scores than those in the non-SVM group. These differences were present before the start of continuous eSCS therapy. *
Across all participants, average daily stimulation use ranged from 5 to 21 h/day, with a mean of 13.7 ± 5.8 h/day. Total eSCS time at their final follow-up visit ranged from 101.2 to 454.5 days, with a mean of 255.3 ± 115.3 days. There were no statistically significant differences in the total amount of stimulation used (
While this study does not involve an intensive rehabilitative therapy component, study personnel collected self-reported information in order to characterize modalities and hours of therapy undergone before and during the study. Six participants reported receiving rehabilitative therapy during acute care as well as general and specialized in-patient rehabilitation. Prior to the study, 57% (4/7) of the participants reported receiving specialized SCI out-patient therapy, 29% (2/7) reported receiving general out-patient therapy, and 43% (3/7) completed therapy at home. After implantation of the epidural stimulator, 43% (3/7) of the participants reported receiving specialized SCI out-patient therapy, 14% (1/7) reported receiving general out-patient therapy, and 43% (3/7) completed therapy at home. Out of the six participants who responded to retrospective surveys, three reported changing their exercise routine post-implantation.
Participants in the SVM group practiced a wide range of exercises spanning from range of motion, aerobic exercises, and general upper- and lower-body strength exercises up to specialized SCI out-patient therapy through adaptive gyms, clinic services with a physical therapist, and even activity-based locomotor exercise programs for core strength, leg strength, and balance. participants from the non-SVM group engaged in a similar range of exercise modalities. Of note, one subject in the non-SVM group participated in rehabilitative therapy that included using a standing frame, exoskeleton, and functional electrical stimulation to aid with stretching sessions. Intensity of rehabilitation therapy varied throughout the overall subject population (
Weekly exercise schedule: reported exercises during study enrollment.
Subject | SVM | Exercises |
1 | No | Stretching and range of motion – 7× week, 30–45 min |
2 | Yes | Activity based locomotor exercise – 3× week, 60 min |
3 | Yes | Upper body exercises – 3× week, 60 min Leg exercises – 3× week, 20 min |
4 | No | Adaptive gym |
5 | Yes | Range of motion, strength training, aerobic exercise, in clinic rehabilitation with a physical therapist – 1× week, 120 min |
6 | Yes | Adaptive gym – 1–2× week, 60 min |
7 | No | Standing frame – 3–4× week, 40 min Exoskeleton – 1× week, 40 min Stretch sessions with electrodes – 2–3× week, 60 min |
Volitional movement without epidural stimulation was observed among the four participants (two females, two males) in the SVM group as early as 3 months post-implantation. In three of the four participants, study personnel cited volitional movement without eSCS during the BMCA. A case report form (CRF) was implemented to document and characterize observed movements with and without stimulation (
Brain Motor Control Assessment (BMCA) documented muscle activation without stimulation
Subject 5 provided a self-report noting right hip adduction, knee flexion/extension, and plantar extension without eSCS at-home during month 13 (
By comparison, movement with stimulation (
In order to characterize study observations, analysis of electrophysiological data from BMCA sessions without stimulation was performed. Sessions in which movement without stimulation was documented demonstrated increases in muscle activity specifically during volitional motor tasks compared to rest during baseline (
BMCA at Subject 2’s follow-up visit 13: surface EMG electrical activity recorded in volts over time for eight bilateral lower extremity muscle groups. Orange traces include left muscle groups and blue traces include right muscle groups. Gray boxes indicate volitional task events signaled by auditory cues. This subsample of EMG recording includes three cues for bilateral hip flexion, right hip flexion, and left hip flexion, respectively. EMG bursts can be seen to be synchronized with the auditory cue followed by silent periods at rest, demonstrating volitional activity.
Observations during BMCA. Differences in volitional power without stimulation (dB) when movement was observed and recorded on case report forms.
A generalized linear model was used to identify the effects of several variables on volitional power (dB) during volitional tasks without stimulation. Longitudinal explanatory variables tested included individual subject, follow-up visit, Modified Ashworth Scores during follow-up visits, and volitional power (dB) with stimulation on. When holding all other variables constant, there was a significant but weak effect of time (
When assessing for the independent effect of time, a trend is apparent as participants progress in follow-up visits (
Average volitional power without stimulation (dB) at each study follow-up visit including all seven participants. It is apparent that Subject 3 demonstrated volitional movement the earliest in the study (follow-up visit 4) and with the greatest magnitude on sEMG, reaching volitional muscle activity 10 times greater (1000%) than that at baseline on follow-up visit 13.
Average volitional power without stimulation (dB) during three follow-up study periods. Study period 1 included follow-up visits 1–5. Study period 2 included follow-up visits 6–9. Study period 3 included follow-up visits 10–13. An improvement over time is apparent only in the SVM group and is statistically significant between study periods 1 and 3 (
Stationary bicycle trials proved complementary to electrophysiological measures as a functional assessment. During one participant’s best bike trial, Subject 3 (
A zero-inflated negative binomial regression for two selected dependent variables, distance traveled and work, was constructed including the following independent variables: individual participants and the interaction between groups (SVM vs. non-SVM) and pedaling effort provided. When holding all other variables constant, there was a significantly strong positive effect on distance traveled (
Functional movement without stimulation: graphs demonstrate differences between Groups (0: non-SVM group and 1: SVM group) and between volitional effort (
Preliminary data from the ESTAND trial suggests that long-term or chronic eSCS can induce plastic changes in chronic, severely injured spinal cords through restored volitional movement without stimulation and without significant intensive rehabilitation. More than half of the first seven patients were observed to exhibit volitional movement without stimulation, which agreed with the more sensitive electrophysiology, and resulted in marked improvements in a functional cycling task.
None of the participants included in this cohort demonstrated traditional signs of discomplete spinal cord injuries before eSCS therapy began (
In an effort to characterize each subject’s propensity for this type of recovery, several descriptive subject characteristics were included such as time since injury, mechanism of injury, injury level, vertebral fracture level, fusion level, and imaging studies. None of these factors were significant in determining if a subject would develop movement without stimulation. Overall, the heterogeneity between participants exemplifies how within the most severe subgroup of the AIS scale, there are no current adequate measures to characterize the functional capacity of the spinal cord. Adequate measures that reflect the degree of preserved and quiescent supraspinal tracts across the spinal cord lesion could allow for phenotyping those responsive to neuromodulation.
In our cohort, spasticity scores before initiating eSCS therapy were slightly greater in the SVM group than in the non-SVM group, which was statistically significant. Moreover, longitudinal spasticity scores before each BMCA session had a significant positive effect on volitional power. When assessing differences among motor-complete SCI participants,
Measures of spinal cord atrophy were also assessed as a marker of injury severity. Only anteroposterior atrophy below the level of injury was found to have a strong effect on volitional power that did not quite meet our criteria for significance. In chronic motor complete spinal cord injuries,
Most importantly, despite the large number of standard characteristics used to describe SCI, all but spasticity proved to be ineffective at predicting the variability in the development of movement without stimulation despite a uniform improvement with active stimulation. In other words, there are no major predictors of the development of movement without stimulation identified thus far that would restrict the potential future use of eSCS as a therapy.
Since the first case of eSCS in SCI aiming to restore volitional movement was reported in 2004 (
Despite heterogeneity in daily time and effort dedicated to physical therapy among all participants, more than half of the participants (SVM group) demonstrated progressive statistically significant improvements in volitional movement in the off stimulation state during the study period. As a result, there were statistically significant improvements in the ability to cycle without assistance, providing the basis for cost-effective home-based therapies to provide incremental improvements in muscle mass, cardio-metabolic risk factors and activities of daily living as well as a platform for activity-dependent plasticity (
While activity-based plasticity is often associated with rehabilitation therapies, there is a possibility that directly increasing volitional movement, increased use, and reliance on these improvements may facilitate more subtle and chronic activity-based plasticity in the sense that directed motor control during normal everyday life drives plastic changes. However, more intensive concurrent rehabilitation outside of the study did not drive further recovery, which is exemplified by the fact that the participant who underwent the most extensive specialized SCI therapy before and during the study did not develop spontaneous movement without stimulation. To our knowledge, this is the first report of eSCS-induced plasticity of volitional movement in the absence of concurrent prescribed intensive locomotor training therapy after motor-complete SCI. While the E-STAND trial remains generalizable by allowing for a wide range of independent therapy, careful data collection of previous therapy regimens may prove useful for assessing further contribution through modeling.
One limitation in this study relates to an undefined off-stimulation time period. As our protocol allows participants to use as much stimulation as they require for their daily activities and comfort. As such, there was no established eSCS-weaning time beyond a minimum of 2 h when testing for off-stimulation activity. An important confounder to consider is whether the reported results might be related to stimulation carry-over effect, described as temporarily persisting changes in spinal cord circuit excitability. This has been described clinically in relation to spasticity modulation in SCI patients as lasting from hours to days (
In this manuscript, the observed movements during the BMCA were not matched to the intended volitional task. Instead, pooling of observed movements was compared to pooled EMG muscle activation during all volitional tasks. In the future, a thorough analysis of the accuracy of muscle activation for each intended joint movement should be performed. The preliminary results at this stage were not sufficiently powered to assess these outcomes. Here we demonstrate the robust evidence of muscle activity magnitude. Muscle activity accuracy will have to be assessed in future larger studies.
Although this is the largest group reported of SCI patients treated with eSCS to restore volitional movement, results should be interpreted with caution due to the small number of participants. In the future, studies with larger cohorts might allow for adequate eSCS therapy phenotyping. In-depth analysis of stimulation usage might point to a dose–response relationship that was not apparent in this study. Furthermore, neurophysiological testing such as somatosensory evoked potentials, electroencephalogram, and transcranial magnetic stimulation motor-evoked potentials might allow for categorizing the effects of chronic eSCS on ascending pathways, cortical representation, and descending pathways, respectively. Correlating these results with high-resolution MRI at enrollment to detect spinal cord area differences might aid in further characterizing the heterogeneity of spinal cord injuries and identifying the degree of preserved pathways that may serve as substrates for recovery (
The datasets presented in this article are not readily available because the ESTAND trial is currently ongoing and datasets retain some identifiable information. A limited dataset may be made available. Requests to access the datasets should be directed to David P. Darrow,
The studies involving human participants were reviewed and approved by the Human Subjects Research Committee, Hennepin Healthcare System. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individuals for the publication of any potentially identifiable images or data included in this article.
DD provided all oversight for the study. DD, APa, AK, APh, US, and TN designed the study. DD, DF, APa, and US performed the critical surgical procedures of this study. CH, AA, SV, DS, DB, and DF performed the study procedures and data collection. DD, TN, NP, and IP analyzed the data. IP wrote the manuscript with support from DD, CH, AA, SV, and DS. All authors provided critical feedback and helped shape the research, analysis, and manuscript, and approved the final manuscript.
US reports having no conflicts of interest relevant to this article. DD, AP, and TN report having several patents related to neuromodulation and are cofounders of a neuromodulation company. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We would like to thank the Minnesota Office of Higher Education SCI/TBI Grant Program (grant number 159800) for the funding to carry out this study and St. Jude/Abbott for a generous device donation.
The Supplementary Material for this article can be found online at:
Summary of means and standard deviations
BMCA video recording of subject 3 with stimulation off. Subjects lie supine with surface electromyography electrodes placed while they perform volitional tasks triggered by auditory cues. At follow up 6, the subject achieves bilateral but predominantly left hip internal rotation and flexion as well as left dorsiflexion of the toes. At follow up 9, he again achieves left hip internal rotation and flexion as well as left dorsiflexion of the toes.
Home video recording of subject 4 demonstrates volitional right knee extension in the absence of stimulation.
Stationary bike video recording of subject 3 at Follow-up visit 6. This trial involves 2 min when subjects intend to pedal with stimulation off. In this particular trial shown, the subject achieved active pedaling for 113 seconds (94% of trial time) for a distance of 268 meters (96% of trial distance).
Brain Motor Control Assessment
epidural spinal cord stimulation
Epidural Stimulation After Neurologic Damage
Modified Ashworth Scale
Spontaneous Volitional Movement.