An implantable system to restore hemodynamic stability after spinal cord injury

Abstract

A spinal cord injury (SCI) causes immediate and sustained hemodynamic instability that threatens neurological recovery and impacts quality of life. Here we establish the clinical burden of chronic hypotensive complications due to SCI in 1,479 participants and expose the ineffective treatment of these complications with conservative measures. To address this clinical burden, we developed a purpose-built implantable system based on biomimetic epidural electrical stimulation (EES) of the spinal cord that immediately triggered robust pressor responses. The system durably reduced the severity of hypotensive complications in people with SCI, removed the necessity for conservative treatments, improved quality of life and enabled superior engagement in activities of daily living. Central to the development of this therapy was the head-to-head demonstration in the same participants that EES must target the last three thoracic segments, and not the lumbosacral segments, to achieve the safe and effective regulation of blood pressure in people with SCI. These findings in 14 participants establish the path to designing a pivotal device trial that will evaluate the safety and efficacy of EES to treat the underappreciated, treatment-resistant hypotensive complications due to SCI.

Fig. 1. Orthostatic hypotension is a medically refractory condition in people with SCI.

a, The prevalence of orthostatic hypotension and management efficacy in 1,479 individuals with tetraplegia (n = 510) from the Rick Hansen Spinal Cord Injury Registry^,. b, Percentage of individuals with tetraplegia experiencing each symptom scored in the ADFSCI survey across various daily activities (n = 107). c, Scheme of tilt-table test and representative blood pressure response. d, Drop in systolic and diastolic blood pressure during a tilt-table test (n = 17). e, ADFSCI total orthostatic hypotension score in individuals with and without clinically defined orthostatic hypotension (n = 4 with no orthostatic hypotension and n = 13 with confirmed orthostatic hypotension; independent samples two-tailed Welch’s t-test: t = 12.84, ***P = 2.37 × 10⁻²). f, Kaplan–Meier plot of exposure status to time, segregated by the presence of clinically defined orthostatic hypotension. Bar graph shows the corresponding time to tilt end. See Source Data Fig. 1 for source data and statistics. Source data

Fig. 2. Validation of the location of the hemodynamic hotspot.

a, Overview of the neuronal architecture recruited EES to trigger pressor responses. b, Anatomical planning of laminotomies for inserting the paddle leads (left) based on the targeted location of the paddle leads (right) over the thoracic and lumbosacral segments. c, Amplitude of systolic pressor response per spinal segment when delivering EES intraoperatively. Dots (n = 5 participants) denote pressure responses to EES applied over specific segments, the location of which was reconstructed postoperatively. d, Bar graph shows averaged changes in systolic blood pressure in response to EES applied over thoracic or lumbar spinal segments (n = 5, paired samples two-tailed t-test; t = 6.46, **P = 3.00 × 10⁻³). e, Systolic blood pressure response measured intraoperatively, while EES was applied over the T11 versus L4 spinal segments for a representative participant. f, Sagittal and coronal reconstructions from a postoperative CT scan. g, Changes in blood pressure during a tilt-table test without EES and with EES applied over either the lower thoracic or lumbosacral spinal segments. h, Drop in systolic blood pressure during a 10-min tilt-table test, and corresponding tilt duration (n = 5, repeated-measures ANOVA with Tukey’s HSD, statistics are provided in Source Data Fig. 2). i, Kaplan–Meier plot of exposure status to time until end of tilt, segregated by the location over which EES was applied (n = 37 tilts; mixed model Cox regression with likelihood ratio test estimate = 36.2, P = 2.48 × 10⁻⁶). See Source Data Fig. 2 and Extended Data Figs. 1–3 for source data and statistics. NS, not significant; HSD, honestly significant difference. Source data

Fig. 3. Validation of the new, purpose-built implantable neurostimulation platform.

a, Scheme of the investigational system, including a clinician controller (tablet), patient programmer (external smartwatch), a main controller (hub), IPG and a Medtronic 5-6-5 SureScan Specify paddle lead. b, Changes in blood pressure from a representative participant during an orthostatic challenge without and with EES applied over the hemodynamic hotspot (EES^BP). c, Average drop in systolic blood pressure during orthostatic challenge without and with EES (n = 4, paired samples two-tailed t-test; t = 4.27, *P = 2.4 × 10⁻²). d, Kaplan–Meier plot of exposure status to time, segregated by the presence or absence of EES (n = 24 tilts; mixed model Cox regression with likelihood ratio test estimate = 15.36, P = 4.00 × 10⁻⁴) and tilt duration (n = 4, paired samples two-tailed t-test; t = 4.69, *P = 1.8 × 10⁻²) without EES and with EES applied over the hemodynamic hotspot. e, Closed-loop control of blood pressure consisting of a PID controller that adjusts the amplitude of EES applied over the hemodynamic hotspot to maintain blood pressure within a range of predefined blood pressure targets. f, Stepwise increase in systolic blood pressure in response to gradual increases of EES amplitude (1 mA per min) for a representative participant. Relationship between EES amplitude and systolic blood pressure (n = 4; black line represents model fit of a mixed effects linear regression R² = 0.97, P = 8.83 × 10⁻⁹). g, Changes in systolic blood pressure during a dynamic orthostatic challenge while EES is delivered continuously (top) or controlled in a closed loop (bottom) to maintain the systolic blood pressure within a target range. h, Error between the predefined blood pressure target (top) and variability of error to target (bottom) when EES was delivered continuously or controlled in closed loop (n = 3/4—trial 3, one participant did not complete assessment due to spasticity). See Source Data Fig. 3 and Extended Data Figs. 4 and 5 for source data and statistics. PID, proportional–integral–derivative. Source data

Fig. 4. Purpose-built paddle lead to target the hemodynamic hotspot.

a, EES applied over multiple segments of the hemodynamic hotspot leads to a superior increase in blood pressure compared to EES applied over the median of a single segment (paired samples, one-tailed t-test, n = 5, t = 2.69, *P = 2.72 × 10⁻²). b, Quantification of the length of the lower thoracic spinal cord in the study participants (n = 10; left), and average length of the lower thoracic spinal cord quantified from published data (right). The coverage of the hemodynamic hotspot by the Medtronic 5-6-5 lead and expected from the newly designed paddle lead are reported. c, Illustrative drawing of the anatomical model implemented in silico to predict the relationships between the lateral position of EES electrodes and the relative recruitment of the targeted dorsal root, contralateral dorsal root and dorsal columns, as shown in the plot. d, Photograph of the newly designed purpose-built paddle lead that combines two parallel columns of eight equally spaced electrodes that aim to cover the entire length of the hemodynamic hotspot while avoiding the recruitment of the dorsal columns. See Source Data Fig. 4 and Extended Data Fig. 6 for source data and statistics. Source data

Fig. 5. Complete purpose-built system to regulate blood pressure.

a, Description of the complete system, including the paddle lead, IPG, communication hub and external smartwatch to operate the various programs of the therapy. b, Postoperative reconstruction of the final position of the paddle lead. c, Changes in blood pressure from a representative participant during an orthostatic challenge without EES and with continuous EES applied over the hemodynamic hotspot. The bar plots report the average drop in systolic blood pressure during the orthostatic challenge and average tilt duration without EES and with EES applied over the hemodynamic hotspot (n = 3). Kaplan–Meier plot of exposure status to time, segregated by the presence or absence of EES (n = 3). d, Changes in diastolic and systolic blood pressure during a dynamic orthostatic challenge while EES is applied continuously (left) or in a closed loop (right) to maintain the systolic blood pressure within a target range (red shaded area) in P11. e, Error between the predefined blood pressure target and variability of error to target when EES was delivered continuously or controlled in closed loop (n = 2/3—trial 3, one participant did not complete the assessment due to spasticity). See Source Data Fig. 5 and Extended Data Fig. 7 for source data and statistics. Source data

Fig. 6. Long-term management of blood pressure improves quality of life.

a, Changes in systolic blood pressure during a formal tilt-table test before implantation (left), without EES after 3–24 months after system implantation (middle) and at the same timepoint with EES (right). n = 11/11—trials 1, 3 and 4; n = 7 measured after at least 6 months of daily management of blood pressure with the system; n = 4 after at least 3 months. b, Bar plot reporting average changes in systolic blood pressure (left) during a format tilt-table test performed as in a (n = 11, repeated-measures ANOVA with Tukey’s HSD). Corresponding Kaplan–Meier plot of exposure status to time sustained in tilt, segregated by the same condition as in a (n = 11, mixed model Cox regression with likelihood ratio test estimate = 36.36; P = 3.0 × 10⁻⁴). c, Changes in systolic blood pressure when delivering EES in a seated position. Each line reports responses from a single participant to EES with default parameters for blood pressure regulation (n = 11/11—trials 1, 3 and 4; n = 9 participants measured after at least 6 months of daily management of blood pressure with the system; n = 2 with at least 3 months). Bar plots report the average change in systolic blood pressure in a seated position that was quantified before and at 2 min after the onset of EES (n = 9 postimplant tilt measured with at least 6 months of EES, n = 2 postimplant tilt measured with at least 3 months of EES use, paired, two-tailed t-test; t = 16.63, ***P = 1.29 × 10⁻⁸). d, Quantification of the total CBF without and with EES using a Doppler spectrum of the internal carotid without and with EES. The bar graph reports the total change in CBF measured in a seated position without and with EES (n = 10/10—trial 1 and 3; paired, two-tailed, t-test; t = −4.32; **P = 1.90 × 10⁻³). e, Bar graph reporting ADFSCI orthostatic hypotension score quantified before implantation of the system and after 6 months of management of blood pressure with the system (n = 12/14—trials 1–4, two participants explanted before 6 months; paired, two-tailed t-test; t = 4.54; ***P = 8.45 × 10⁻⁴). Symptoms are color-coded by the percentage of participants who experience each symptom. f, Bar graph reporting the performance (n = 8/8—trials 3 and 4, paired, two-tailed t-test; t = −3.92; **P = 5.7 × 10⁻³) and satisfaction (n = 8/8—trials 3 and 4, paired, two-tailed t-test; t = −5.54; ***P = 8.7 × 10⁻⁴) scores of the COPM relative to baseline goals set by occupational therapists before the implantation of the system and after 6 months of management of blood pressure with the system. g, Bar graph reporting scores on the MOS-S—Sleep Problem Index II measured before implantation of the system and after 6 months of management of blood pressure with the system (n = 12/14—trials 1–4, two participants explanted before 6 months; paired, two-tailed t-test; t = 2.58; *P = 2.52 × 10⁻²). h, Bar graph reporting NBDS before implantation of the system and after 6 months of management of blood pressure with the system (n = 11/11—trials 1–3; paired samples t-test; t = 2.23; *P = 4.95 × 10⁻²). i, Usage of EES programs to improve blood pressure instability during upright rehabilitation and maintain trunk stability during recreation activities such as sit-skiing. See Source Data Fig. 6 and Extended Data Figs. 9 and 10 for source data and statistics. Source data

Extended Data Fig. 1. Mapping of the optimal location to modulate blood pressure with EES.

a, Timeline of clinical trial 1 (STIMO-HEMO) conducted in Lausanne, Switzerland and trial 2 (HEMO) in Calgary, Canada. b, Changes in systolic blood pressure during a formal 10-minute tilt-table test to verify that the six participants met the criteria for confirmed orthostatic hypotension. The bar graph reports the average drop in systolic blood pressure at or within (if the test had to be aborted before) 3 min of verticalization at a 70-degree tilt position. c, As in b, for diastolic blood pressure. d, As in b, for mean arterial pressure. e, As in b, for heart rate. f, Bar graph reporting the average duration of the tilt-table test shown in b–e. Each point denotes a single tilt for each participant (n = 6, left), with the corresponding Kaplan–Meier plot of exposure status to time sustained in tilt (right). g, Schematic of the investigational device used to deliver EES in both trials, which includes 2 Medtronic Intellis Implantable Pulse Generators (IPG), 2 Medtronic Specify SureScan 5-6-5 Leads and Medtronic clinical and patient programmers for home use. h, A personalized anatomical model of the spine is elaborated for each participant based on high-resolution magnetic resonance imaging and computed tomography. This model guides the determination of the desired lead placement and spinal entry levels (location of laminectomy). The final location of the leads is optimized intraoperatively based on the monitoring of electromyographic signals from key trunk and leg muscles (bottom right). i, The locations of each electrode are captured with postoperative computed imaging, which are then injected into the personalized anatomical model to visualize the final placement of both leads. j, Final location of the electrodes in both trials (n = 6). k, Changes in systolic blood pressure in response to EES applied over the segments of the lower thoracic and lumbosacral spinal cord during the intraoperative mapping for each participant. The location of the segments was determined postoperatively. Due to time constraints and electrode locations, only a subset of segments could be tested for each participant. l, Bar plots reporting the average change in diastolic blood pressure (left; n = 5, paired samples two-tailed t-test; t = 6.17; p = 3.50e-03) and mean arterial pressure (right; n = 5, paired samples two-tailed t-test; t = 7.55; p = 1.70e-03) in response to EES applied over the lower thoracic versus lumbosacral spinal segments during the intraoperative mapping. See Source Data Fig. 2 and Extended Data Figs. 1–3 for source data and statistics. Source data

Extended Data Fig. 2. Personalized anatomical models of the spine.

Step 1—acquisition of high-definition magnetic resonance imaging (MRI) images of the relevant region of the spine with MRI sequences that have been optimized to augment the contrast between the dorsal root and cerebrospinal fluid. Step 2—accurate detection of the centerline of the spinal cord, even in areas with low signal or artifacts. Creation of a 35 mm cylindric mask around the centerline and cropping of the image into boxes along the rostrocaudal axis. Step 3—a neural network trained on segmented data from 12 healthy volunteers is applied on the MRI acquisitions to detect the epidural fat, cerebrospinal fluid, white matter and spinal root tissues. The segmented boxes are then merged, and manual corrections are performed. Step 4—a mask from the segmented epidural fat is used to refine the segmentation, with an option for manual adjustments. Step 5—the contour of ventral and dorsal roots are traced using the MRI and segmented tissues as guides, which enables the labeling of roots as they exit the vertebrae. Step 6—the trajectories of the spinal roots are trimmed within white matter tissues to 1 mm. These endpoints are used to determine the spinal segments and trajectories of the roots. This procedure is robust to the presence of scoliosis. Step 7—the trajectory of each spinal root is subdivided into 5 rootlets to generate anatomically realistic tridimensional models of the rootlets. Step 8—acquisition of low-resolution MRIs with enhanced bone-tissue contrast. Step 9—segmentation of bones, disks, and spinal canal using a neural network trained on low-quality data. Step 10—removal of incomplete vertebrae/disks, discarding spinal canal segmentation. Step 11—labeling of intervertebral disks for vertebrae identification to facilitate the coregistration of CT and MRI acquisitions. Step 12—acquisition of high-definition computerized tomography scan. Step 13—segmentation of individual vertebrae from the computerized tomography scan. Step 14—alignment of computerized tomography and MRI data using labeled vertebrae and transformation with a two-step function. Step 15—personalized anatomical model of the spine is generated based on the surfaces of the segmented tissues extracted from the MRI images. Preoperative and postoperative computerized tomography scans are used to update the model with the implanted system. Source data

Extended Data Fig. 3. The hemodynamic hotspot to regulate blood pressure with EES is located in the low thoracic spinal cord, not the lumbosacral spinal cord.

a, Changes in systolic blood pressure during a formal tilt-table test performed without EES or with EES applied over the lower thoracic segments or the lumbosacral segments during postoperative evaluations. Each trace corresponds to a representative test under each experimental condition for the 5 participants. b, As in a, for diastolic blood pressure. c, As in a, for mean arterial pressure. d, As in a, for heart rate. e, Bar graphs reporting the average drop in diastolic blood pressure at or within (if aborted) 3 min during a formal tilt-table test without EES, or with EES applied over the lower thoracic segments or the lumbosacral segments. Three tilts were obtained for P1–P3 in each condition; 2 tilts were obtained for P4 without EES and 1 tilt in each condition with EES; 2 tilts were obtained for P5 in each condition (n = 5, repeated-measures, two-tailed ANOVA with Tukey’s HSD; f = 19.6; p = 8.3e-04). f, As in e, for mean arterial pressure (n = 5, repeated measures, two-tailed ANOVA with Tukey’s HSD; f = 19.27; p = 8.7e-04). g, As in e, for heart rate (n = 5, repeated-measures, two-tailed ANOVA with Tukey’s HSD; f = 0.29; p = 0.76). See Source Data Fig. 2 and Extended Data Figs. 1–3 for source data and statistics. Source data

Extended Data Fig. 4. Validation of the purpose-built neurostimulation platform.

a, Timeline of clinical trial 3 (HemON) clinical trial conducted in Lausanne, Switzerland in 4 participants. b, Changes in systolic blood pressure during a formal 10-min tilt-table test to verify that the four participants met the criteria for confirmed orthostatic hypotension (red dotted line). The bar graph reports the average drop in systolic blood pressure at or within (if the test had to be aborted before) 3 min of verticalization at a 70-degree tilt position. c, As in b, for diastolic blood pressure. d, As in b, for mean arterial pressure. e, As in b, for heart rate. f, Bar graph reporting the average duration of the tilt-table test shown in b–e. Each point denotes a single tilt for each of the 4 participants. g, Schematic illustration of the investigational device used to deliver EES, including the ONWARD ARC^IM Implantable Pulse Generator (IPG), the Medtronic Specify SureScan 5-6-5, the ARC^IM Hub for therapy control and IPG charging and the ONWARD ARC^IM clinician and personal programmer for home use. h, Final location of the electrodes with respect to the hemodynamic hotspot (red) for the 4 participants. i, Changes in systolic blood pressure when delivering EES targeting the hemodynamic hotspot (EES^BP) in a seated position for each of the 4 participants. The bar graphs report average changes in systolic blood pressure (paired, two-tailed, t-test; t = 6.75, **p = 6.65e-03). j, As in i, for diastolic blood pressure (paired, two-tailed t-test; t = 7.56, **p = 4.81e-03). k, As in i, for mean arterial pressure (paired, two-tailed t-test; t = 7.12, **p = 5.70e-03). l, As in i, for heart rate (paired, two-tailed t-test; t = 3.10, p = 5.31e-02). m, Changes in systolic blood pressure without (gray) or with EES^BP (red) during a formal tilt-table test. n, As in m, for diastolic blood pressure. o, As in m, for mean arterial pressure. p, As in m, for heart rate. q, Bar plots reporting the average change in diastolic blood pressure during a formal tilt-table test without and with EES^BP (paired, two-tailed t-test; t = 3.84; *p = 3.1e-02). r, As in q, for mean arterial pressure (paired, two-tailed t-test; t = 4.00; *p = 2.8e-02). s, As in q, for heart rate (paired, two-tailed t-test; t = 1.63; p = 0.20). See Source Data Fig. 3 and Extended Data Figs. 4 and 5 for source data and statistics. Source data

Extended Data Fig. 5. Closed-loop control of EES to stabilize blood pressure despite constant-changing orthostatic challenges.

a, Changes in systolic blood pressure in response to graded increase in the amplitude of EES by 1 mA increment every minute. b, Schematic view of the three different paradigms used to control EES^BP during dynamic changes of the angle of the tilt-table (left). Changes in blood pressure while EES^BP is delivered continuously (top), with closed-loop control of EES^BP based on the amplitude of the tilt measured with an IMU (middle) and with closed-loop control of EES^BP based on continuous monitoring of systolic blood pressure (bottom) for the 3 participants. c, Bar plots reporting the average error to target (left) and variability of the error to target (right) for the 3 experimental conditions described in b. The target was based on a predefined level of systolic blood pressure. See Source Data Fig. 3 and Extended Data Figs. 4 and 5 for source data and statistics. Source data

Extended Data Fig. 6. Informed design of the purpose-built paddle lead.

a, Personalized anatomical models of the spinal cord updated with the final placement of the paddle lead for the 10 participants who were implanted with the Medtronic SureScan Specify 5-6-5 paddle lead (Supplementary Table 2). The hemodynamic hotspot is represented in red, thus informing the relative coverage of this hotspot with this specific lead. b, Examples of undesired responses in leg muscles when EES was applied with over electrode configurations that likely recruited the dorsal columns and thus elicited muscle activity via antidromic afferent volleys evoked on ascending afferent fibers. c, Quantification of the location of the dorsal root entry zones and of the width of the spinal cord based on high-resolution magnetic resonance imaging acquired in the 10 participants shown in a. d, Computer simulations to predict the relative recruitment of the targeted dorsal root, dorsal columns, and contralateral dorsal root based on the location of the electrode along the lateral directions. A canonical model of the lower thoracic spinal cord was implemented in silico for these simulations. See Source Data Fig. 4 and Extended Data Fig. 6 for source data and statistics. Source data

Extended Data Fig. 7. Validation of the complete purpose-built system to treat hemodynamic instability.

a, Timeline of clinical trial 3 (HemON) amended for the new paddle lead conducted in Lausanne, Switzerland in 3 participants (Supplementary Table 2). b, Changes in systolic blood pressure during a formal 10-minute tilt-table test to verify that the three participants met the criteria for confirmed orthostatic hypotension (red dotted line). The bar graph reports the average drop in systolic blood pressure at or within (if the test had to be aborted before) 3 min of verticalization at a 70-degree tilt position. c, As in b, for diastolic blood pressure. d, As in b, for mean arterial pressure. e, As in b, for heart rate. f, Bar graph reporting the average duration in tilt shown in b–d. Each point represents a single tilt-table test of a single participant (n = 3). g, Computed tomography and structural magnetic resonance imaging acquired preoperatively to elaborate personalized models of the spine for each participant. h, Anatomical model, preoperative planning and postoperative reconstruction of the final placement of the lead for each participant (n = 3). i, Pressor responses elicited on the first day after the surgical implantation of the system when delivering EES^BP in a supine position (P11, P12—arterial line; P13—finapres). j, Bar graph reporting the average change in systolic blood pressure without and with EES^BP. k, As in j, for diastolic blood pressure. l, As in j, for the mean arterial pressure. m, Change in systolic blood pressure during a formal 10-minute tilt-table test without and with EES^BP. Each line corresponds to a single test. n, As in m, for diastolic blood pressure. o, As in m, for mean arterial pressure. p, As in m, for heart rate. q, Bar graph reporting the average change in systolic blood pressure without and with EES^BP at or within (if the test had to be aborted before) 3 min of verticalization at a 70-degree tilt position. r, As in q, for diastolic blood pressure. s, As in q, for mean arterial pressure. t, As in q, for heart rate. u, Changes in systolic blood pressure with increases of EES^BP of 1 mA every 1 min (P11, P13). These recordings could not be collected in P12. v, Relationship between the amplitude of EES^BP and the modulation of systolic blood pressure for P11 and P13. See Source Data Fig. 5 and Extended Data Fig. 7 for source data and statistics. Source data

Extended Data Fig. 8. Independent validation of the complete system.

a, Timeline of clinical trial 4 (HemON-NL) conducted in Nijmegen, Netherlands. b, Changes in systolic blood pressure during a formal 10-minute tilt-table test to verify that the four participants met the criteria for confirmed orthostatic hypotension (red dotted line). The bar graph reports the average drop in systolic blood pressure at or within (if the test had to be aborted before) 3 min of verticalization at a 70-degree tilt position. c, As in b, for diastolic blood pressure. d, As in b, for mean arterial pressure. e, As in b, for heart rate. f, Bar graph reporting the average duration in tilt shown in b–d. g, Pressor response recorded intraoperatively using an arterial line. h, Final placement of the paddle lead represented into the personalized anatomical model of the spinal cord. i, Pressor responses recorded from a seated position during the first application of EES^BP after implantation of the system. The scheme of the paddle lead indicates the configuration of the electrodes, which was determined based on the location of the electrodes with respect to the targeted dorsal root entry zones. Bar graphs report the change in systolic blood pressure, diastolic blood pressure, and mean arterial pressure quantified 2 min after the application of EES^BP. j, Change in systolic blood pressure without and with EES^BP during a formal 10-minute tilt-table test. Each line corresponds to a single test. The bar graph reports the average drop in systolic blood pressure at or within (if the test had to be aborted before) 3 min of verticalization at a 70-degree tilt position. k, As in j, for diastolic blood pressure. l, As in j, for mean arterial pressure. m, As in j, for heart rate. n, Bar graph reporting the average duration in the tilt-table test shown in j–m. o, Changes in systolic blood pressure with increases of EES^BP of 1 mA every 1 min, and relationship between the amplitude of EES^BP and the modulation of systolic blood pressure (R² = 0.96, p-value = 1.90e-02). p, Changes in diastolic and systolic blood pressure during a dynamic orthostatic challenge while EES^BP is applied continuously (left) or in closed loop (right). q, Bar plots reporting the average error to target (left) and variability of the error to target (right) while EES^BP is applied continuously (left) or in closed loop (right). See Source Data Extended Data Fig. 8 for source data and statistics. Source data

Extended Data Fig. 9. Long-term efficacy of the system to regulate hemodynamic instability.

a, Evaluation of pressor responses from a seated position in a wheelchair. b, Changes in systolic blood pressure when delivering EES^BP from a seated position (a). Each line corresponds to a response obtained in one of the 11 participants using the default EES^BP program after 3 months (n = 2) or at least 6 months after implantation of the system. The black line corresponds to the average response for all participants. The bar plots report the average change in systolic blood pressure in seated position before and 2 min after the onset of EES^BP (n = 11/11—trial 1, 3, 4; n = 9 measured at least 6 months after implantation, n = 2 measured with at least 3 months after implantation, paired, two-tailed t-test; t = 16.63; ***p = 1.29e-08). c, As in b, for diastolic blood pressure (paired, two-tailed t-test; t = 14.01; ***p = 6.75e-08). d, As in b, for mean arterial pressure (paired, two-tailed t-test; t = 15.77; ***p = 2.16e-08). e, As in b, for heart rate (paired, two-tailed t-test; t = 3.39; **p = 6.9e-03). f, Timeline for long-term efficacy assessments, and time windows during which blood samples were taken during a formal 10-minute tilt-table test. g, Change in systolic blood pressure during a formal 10-minute tilt-table test conducted before implantation of the system (left), at least 3 months after implantation of the system but with EES^BP turned off (middle) and at the same timepoint with EES^BP turned on (right; n = 11/11—trial 1, 3, 4; n = 7 measured at least 6 months after implantation; n = 4 measured at least 3 months after implantation). Bar graph reports changes in systolic blood pressure during the formal tilt-table test. Each dot corresponds to a single test (n = 11/11—trial 1, 3, 4; repeated-measures ANOVA with Tukey’s HSD). h, As in g, for diastolic blood pressure (n = 11; repeated-measures ANOVA with Tukey’s HSD). i, As in g, for mean arterial pressure (n = 11; repeated-measures ANOVA with Tukey’s HSD). j, As in g, for heart rate (n = 11; repeated-measures ANOVA with Tukey’s HSD). k, Bar plots reporting the average duration in the tilt-table test shown in g–j. Each dot corresponds to a single test per participant (n = 11; repeated-measures ANOVA with Tukey’s HSD). l, Bar plots reporting changes in the concentration of endogenous norepinephrine in the blood between the supine position (T0) and after 2 min of verticalization at 70 deg (T2) without and with EES^BP (n = 10/13—trials 1,2 and 3; 3 participants could not be evaluated as described in methods, paired samples t-test, t = 5.49, ***p-value = 3.90e-04). m, Schematic of the experimental setup and algorithmic framework to control EES^BP in closed loop to maintain systolic blood pressure within a user-defined target range. n, Changes in systolic and diastolic blood pressure during repeated transitions from tilts maintained at 20 or 70 degrees without EES^BP and with closed-loop control of EES^BP. o, Bar plots reporting the average drop in systolic blood pressure during the tilts (n = 5, paired, two-tailed t-test, t = 5.78, **p = 4.44e-03 average maximum drop during tilt down; t = 4.81, **p = 8.57e-03 average maximum drop during tilt up), and average variability of error to target without EES^BP and with closed-loop control of EES^BP (n = 5, paired samples t-test, t = 5.75, **p = 4.55e-03 average variability during tilt down; t = 10.39, ***p = 4.84e-04 average variability during tilt up). Each dot corresponds to the average values for a complete sequence per participant. See Source Data Fig. 6 and Extended Data Figs. 9 and 10 for source data and statistics. Source data

Extended Data Fig. 10. Long-term usability and efficacy of the system.

a, Bar plots reporting the average daily usage of the system per participant (left) and the usage of the system throughout the day (right) for all participants implanted with at least part of the system (n = 8/8—trials 3–4). b, Bar plots reporting changes in blood pressure measured 30 min before and every 2–5 min from 30 to 60 min after a controlled food intake without and with EES^BP for participants presenting with postprandial hypotension (systolic, n = 7, paired samples t-test, t = 4.13, **p-value = 6.12e-03; diastolic blood pressure, n = 7, paired, two-tailed, t-test, t = 2.66, *p-value = 3.78e-02; mean arterial blood pressure, n = 7, paired samples t-test, t = 3.37, *p-value = 1.50e-02—trials 1–3). c, Bar plots reporting the score on the system usability scale during formal usability testing of the 7 participants implanted with at least the implanted pulse generator of the complete system (n = 7/7—trial 3). d, Bar plots reporting percentage change of the 12 participants who took midodrine medication (left) and supportive garments (right) including compressive belts and/or stockings, before implantation of the system and after 6 months of home use (left; n = 12/14—trials 1–4, 2 participants in trial 2 explanted before 6 months). e, Bar plots reporting spasticity score for the upper limb (left) and lower limb (right) without medication based on the modified Ashworth scale, quantified without and with programs to reduce muscle spasms (EES^Spasticity; n = 10/10—trials 1 and 3; left: paired, two-tailed t-test; t = 2.30; *p-value = 4.80e-02, right: paired two-tailed t-test; t = 3.30; **p-value = 9.31e-03). f, Percentage of use of EES for various exercises measured over 1 month of rehabilitation for the seven participants who followed a personalized training program (n = 8/8—trials 3–4). g, Bar plot reporting changes in percentage of thoracic spine curvature without and with EES targeting the trunk musculature (EES^Trunk; n = 7/10—trials 1 and 3, 3 participants not evaluated). h, Bar plot reporting the peak expiratory flow during a coughing maneuver without and with EES targeting trunk timed with coughing volition (EES^Coughing; n = 9/10—trials 1 and 3, 1 incomplete assessment due to adverse event; paired, two-tailed, t-test; t = 3.63; **p-value = 6.70e-03). i, Changes in splanchnic volume with and without EES^BP (left), and bar plots (right) reporting changes in splanchnic volume for 2 participants from the trial conducted in Calgary (n = 2/3, trial 2, 1 participant explanted before assessment). j, Muscle sympathetic nerve activity captured in the spike waveform, neurogram, integrated neurogram and corresponding blood pressure without and with EES^BP for P4 (left). Bar plots reporting changes in neuronal spike frequency without and with EES^BP for P4 (right)—(1/3, trial 2, 2 participants explanted before assessment). k, Representative participant’s quotes extracted from semi-structured interviews expliciting changes in quality of life with the therapy. See Source Data Fig. 6 and Extended Data Figs. 9 and 10 for source data and statistics. Source data