Noninvasive spinal stimulation safely enables upright posture in children with spinal cord injury

Abstract

In children with spinal cord injury (SCI), scoliosis due to trunk muscle paralysis frequently requires surgical treatment. Transcutaneous spinal stimulation enables trunk stability in adults with SCI and may pose a non-invasive preventative therapeutic alternative. This non-randomized, non-blinded pilot clinical trial (NCT03975634) determined the safety and efficacy of transcutaneous spinal stimulation to enable upright sitting posture in 8 children with trunk control impairment due to acquired SCI using within-subject repeated measures study design. Primary safety and efficacy outcomes (pain, hemodynamics stability, skin irritation, trunk kinematics) and secondary outcomes (center of pressure displacement, compliance rate) were assessed within the pre-specified endpoints. One participant did not complete the study due to pain with stimulation on the first day. One episode of autonomic dysreflexia during stimulation was recorded. Following hemodynamic normalization, the participant completed the study. Overall, spinal stimulation was well-tolerated and enabled upright sitting posture in 7 out of the 8 participants.

Conflict of interest statement

Dr. Yury Gerasimenko has a shareholder interest in NeuroRecovery Technologies and Cosyma. He holds certain inventorship rights on intellectual property licensed by the regents of the University of California to NeuroRecovery Technologies and its subsidiaries. Dr. Anastasia Keller, Dr. Goutam Singh, Joel Sommerfel, Molly King, Parth Parikh, Dr. Beatrice Ugiliweneza, Dr. Jessica D’Amico, and Dr. Andrea L. Behrman declares no competing interests.

© 2021. The Author(s).

Figures

Fig. 1. Hemodynamic parameters in response to acute transcutaneous spinal cord stimulation (scTs) in children with spinal cord injury (SCI).

Systolic (a) and diastolic (b) blood pressure (millimeter of Mercury, mmHg) and heart rate (c) (beats/minute) measurements during experiments at baseline (BL, 1 trial), with scTS at T11, L1 (and C5 when tolerated, 1 trial) and at the end of experiment (Post, 1 trial) graphed as means + standard deviation over 3 days of assessments for each participant (n = 7). No significant changes were observed in the systolic (F2,12 = 1.98, p = 0.18), diastolic (F2,12 = 2.27, p = 0.1462) blood pressures or heart rate (F2,12 = 0.2, p = 0.82) between the designated assessment time points at baseline, with scTS on and post experiment.

Fig. 2. Sitting posture during transcutaneous spinal cord stimulation (scTS) vs. passive pelvic tilt in children with SCI.

Segmental trunk kinematics during scTS optimization for representative participants P14 (a) and P23 (e). Anteroposterior and mediolateral center of pressure displacements (millimeters, (mm)) were recorded concomitant with kinematics P14 (d), P23 (h). Manual pulse indicates the increase of stimulation intensity in 10 milliamp (mA) increments. Trunk kinematics during passive pelvic tilt was performed by a physical therapist without scTS while participants were seated relaxed, P14 (b) and P23 (f). Black curly bracket indicates the attempt to shift pelvis from posterior tilt toward neutral position, gray curly bracket indicates participants response of falling forward. Additional support at the anterior aspect of the shoulders was then provided by another therapist during passive pelvic tilt for P14 (c) and P23 (g). The participants were instructed to maintain upright posture after the shoulder support was removed with just the pelvic support. The red curly bracket indicates the participant’s P14 attempt to stay upright (c). Participant P23 was not able to maintain balance once shoulder support was removed (g). Gray arrow points to the perturbation in trunk kinematics at the initial fall. The participant was spotted (black arrow) and repositioned to upright posture which P23 could not maintain, falling forward again when shoulder support was withdrawn.

Fig. 3. Acute effects of lumbosacral transcutaneous spinal cord stimulation on segmental trunk extension in children with SCI.

Box plots of angles at each measured trunk segment averaged over 10 s of the participants’ volitional attempt (VA) to sit upright (white boxes) (average of 3 trials), prior to stimulation, baseline (BL) resting sitting (gray boxes) or during 10 s of sitting with scTS at T11 (a) and L1 (b) at the upright posture-inducing, scTS intensities optimized for each individual (blue boxes) (n = 7 participants for each stimulation site, 1 trial at each stimulation site on the third day). Change in anteroposterior and mediolateral center of pressure displacement (COP, millimeters, mm) during scTS at T11 (c) and L1 (d) relative to baseline (n = 6 for each stimulation site, the missing data point from one of the experiments occurred due to the loss of signal from the force plate). The centerline represents the group median, the left, and right box bounds represent 25th and 75th interquartile range (IQR), respectively. Box whiskers represent 1.5 times the IQR. The overlaying dots represent individual data points. Black dots are outlier points that lie outside of 1.5 times the IQR. The dotted line at 0° represents neutral vs. extended (−) or flexed (+) trunk position. Mixed linear regression models were used to assess the overall differences in trunk angles and COP changes across the timepoints of baseline sitting, volitional attempt, compared to stimulation at T11 (F16,140 = 14.41, p < 0.0001) and L1 (F16,140 = 7.75, p < 0.0001), followed by Tukey’s post hoc t test. *denotes significance for L5S1: BL vs. T11 scTS, p = 0.03, Cohen’s d = 0.97; BL vs. L1 scTS, p = 0.006, Cohen’s d = 1.18; VA vs. L1 scTS, p = 0.03, Cohen’s d = 0.97; for PelvisT8: BL vs. T11 scTS, p < 0.0001, Cohen’s d = 2.1; VA vs. T11 scTS, p < 0.0001, Cohen’s d = 2.4; BL vs. L1 scTS, p < 0.0001, Cohen’s d = 2.35; VA vs. L1 scTS, p < 0.0001, Cohen’s d = 2.76; for T8Head: BL(T11) vs. VA, p = 0.03, Cohen’s d = 0.95; BL vs. T11 scTS, p = 0.0027, Cohen’s d = 1.28; VA vs. T11 scTS, p < 0.0001, Cohen’s d = 2.23; BL (L1) vs. VA, p = 0.0002, Cohen’s d = 1.53; BL vs. L1 scTS, p < 0.002, Cohen’s d = 1.33; VA vs. L1 scTS, p < 0.0001, Cohen’s d = 2.86; for anteroposterior COP: change from BL vs. T11 scTS, p = 0.0008, Cohen’s d = 1.53, change from BL vs. L1 scTS, p < 0.0001, Cohen’s d = 2.95; VA vs. L1 scTS, p = 0.0001, Cohen’s d = 1.74; for mediolateral COP: VA vs. L1 scTS, p = 0.0023, Cohen’s d = 1.4.