Two-Year Retention of Benefits After Paired Vagus Nerve Stimulation in Stroke

Abstract

Objectives

Vagus nerve stimulation (VNS) paired with rehabilitation improves short-term upper extremity (UE) function in people with chronic ischemic stroke. Because sustained improvements are needed for people with stroke, we examined the long-term effects of Paired VNS on impairment, activity, participation, and quality of life.

Methods

This is a post hoc analysis of 2-year data of the triple-blind, sham-controlled, randomized VNS-REHAB pivotal trial. After receiving the full therapeutic protocol with either real (Active) or sham (Control) VNS, participants continued self-activated active stimulation for at least 2 years. The Fugl-Meyer Assessment-Upper Extremity (FMA-UE), Wolf Motor Function Test (WMFT), participation, and quality of life outcomes were assessed at multiple time points.

Results

Pooled 2-year outcome data were available from 49 participants. At the 2-year point, FMA-UE scores improved 7.51 points (95% CI 5.80–9.22; p < 0.001), and WFMT scores improved 0.63 points (0.50–0.75; p < 0.001) from baseline. There were significant improvements from baseline in 5 of 7 participation and quality of life measures at 2 years. Improvements were retained at 3 years among a subset of participants (n = 16).

Discussion

The sustained improvements in UE impairment and function for at least 2 years make this an important treatment option for a defined subset of individuals with chronic arm and hand deficits after ischemic stroke.

Trial Registration Information

ClinicalTrials.gov/study/NCT03131960.

Introduction

Stroke is a chronic condition that often results in persistent loss of function and quality of life for most survivors.^1,2 As individuals are living longer with poststroke disability, treatments with durable benefits are essential to support sustained functional independence and well-being.^3,4

We previously reported the safety and efficacy of neuromodulation using implanted paired vagus nerve stimulation (VNS) (paired VNS) on upper extremity (UE) impairment and function in people with ischemic stroke in the pivotal sham-controlled, triple-blind, randomized VNS-REHAB trial.⁵ Positive outcomes from the study led to the approval of paired VNS (Vivistim Paired VNS System, MicroTransponder Inc., Austin, TX) to treat moderate-to-severe UE motor deficits associated with chronic ischemic stroke. Subsequently, we showed that the improvements in impairment, activity, patient-reported outcomes, and quality of life measures were maintained at 1 year⁶ Here, we present the long-term outcomes beyond 1 year, from available VNS-REHAB trial participants.

Methods

The design of the VNS-REHAB trial, a triple-blinded, sham-controlled pivotal study, has been described previously including the original statistical analysis plan and protocol.^5,7 The study adhered to the principles of the Declaration of Helsinki and was conducted across 19 sites in the United States and the United Kingdom.

Participants

Eligible participants were aged 22–80 years, had a unilateral supratentorial ischemic stroke, a Fugl-Meyer Assessment-Upper Extremity (FMA-UE) score between 20 and 50/66 (higher is better), and some active wrist/finger movement. Complete inclusion and exclusion criteria appear in eMethods.⁷

Intervention

In the VNS-REHAB trial, all participants (n = 108; 53 VNS, 55 control) were implanted with the VNS device.⁵ Participants were randomly assigned to either active or sham VNS and completed 6 weeks (18 sessions) of in-clinic UE rehabilitation,⁷ which was high repetition, individualized, task-based, functional, and progressive. This was followed by 3 months of daily 30-minute therapist-designed, individualized, self-activated, home-based therapy (eMethods). All a priori primary and secondary data from the blinded randomized study phase have been previously reported.⁵ Thereafter, participants in the VNS group continued with the long-term phase, performing self-activated therapy with active VNS. Participants who were initially in the control group entered a cross-over phase, where they completed a course of active VNS paired with the same in-clinic rehabilitation protocol and self-activated period with the same assessment schedule, followed by the long-term phase.⁶ Here, we report outcomes 2 years or longer after completion of the VNS intervention (Figure 1).

Figure 1. Study Timeline.

Participants were randomly assigned to Active VNS (top row—red) or Control (bottom row—blue solid) groups and completed 6 weeks of active VNS with rehabilitation. Assessments occurred at Day 1 and Day 90 after completion of the in-clinic therapy. Control group participants crossed over to receive active VNS + rehabilitation during the cross-over phase after the Day 90 assessment (bottom row-blue dashed).⁵ The long-term phase began for all participants after they completed the active VNS Day 90 assessment, which for the VNS group was approximately 135 days after randomization, and for the control group, it was approximately 279 days after randomization. Participants in both groups were instructed to perform self-activated VNS throughout the long-term phase. For both groups, assessments were conducted at 1⁶ and 2 years after the completion of active VNS; 3-year data were available for a subset of individuals and are reported in the supplementary material. All data for the long-term phase are pooled (green). *Results previously reported.^5,6 VNS = Vagus nerve stimulation.

Outcomes

We examined changes from baseline to 2 and 3 years post–Active Paired VNS for the FMA-UE, Wolf Motor Function Test-Functional Ability Scale (WMFT-FAS), and patient-reported outcome measures (PROMs) including: Stroke Impact Scale-Activity of Daily Living and Hand subscales (SIS-ADL, SIS-Hand), Motor Activity Log-Amount of Use and Quality of Movement scales (MAL-AOU, MAL-QOM), Stroke Specific-Quality of Life scale (SS-QOL), EQ-5D, and Beck Depression Inventory (BDI).

Statistical Analysis

Data were analyzed for the subset of participants with 2-year outcomes (n = 49). Data were pooled from both randomized groups, as all participants had received both in-clinic and self-activated active Paired VNS during the long-term phase. All changes are reported relative to the original baseline unless noted otherwise.

A linear mixed-effects regression model with random subject-specific intercepts was used to analyze the outcomes for FMA-UE, Wolf Motor Function Test (WMFT), and PROMs. This model is suitable for analyzing repeated measures over time and for handling missing data. All models controlled for age, sex, time poststroke, and side of paresis. Given the number of PROMs (n = 7), p values were adjusted for multiple comparisons using a Bonferroni correction. Clinical meaningfulness was examined by constructing participant response profiles based on a minimal clinically important difference (MCID)–level improvement, as previously performed for 1-year data (eMethods).⁶ Analyses were conducted in RStudio (version 2025.05.1).⁸ Changes from baseline were reported with 95% confidence intervals and p values. p Values < 0.05 were considered statistically significant.

Standard Protocol Approvals, Registrations, and Participant Consents

Institutional review boards (IRB) at all sites approved the study (Coordinating IRB: WCG Clinical, Inc. (WCG–IRB); Sponsor Protocol#: MT-St-03 (VNS-REHAB); Protocol#: 20171052), and regulatory clearance was obtained. The trial was registered at ClinicalTrials.gov (NCT03131960). The study protocol and statistical analysis plan were published and are freely available as open access.⁵ All participants provided written informed consent.

Data Availability

Data sharing requires a signed data access agreement with specific outside funding explicitly stated to access the database without any support from investigators or MicroTransponder. Requests should be sent to VNSdatarequest@gmail.com.

Results

Participants

Forty-nine of the original 108 participants completed the 2-year follow-up (Figure 2). Baseline characteristics and demographics are presented in eTable 1. Most missing data were due to SARS-CoV-19 pandemic restrictions, as most long-term follow-up visits occurred between 2020 and 2022.⁶ There were no differences in demographics or baseline characteristics of individuals included and excluded from this analysis (eTable 2).

Figure 2. CONSORT (Consolidated Standards of Reporting Trials) Diagram.

Flow of participants through the blinded, crossover (Control group only, Day 1_cross and Day 90_cross), and unblinded, long-term follow-up phases (pooled: Active and Control groups) of the VNS-REHAB trial. Forty-nine participants attended the 2-year follow-up assessment and were analyzed in this report. For analyses examining 3-year follow-up data, only individuals who previously attended the 2-year assessment were included; thus, 16 of the 24 possible participants were assessed for all 3-year results. The remaining 8 participants were excluded from the analyses reported here. ITT = indicates intent to treat; and VNS = vagus nerve stimulation.

Changes in UE and Participant-Reported Outcomes at 2 Years

At the 2-year assessment, the average FMA-UE and WMFT-FAS scores improved by 7.51 points (95% CI 5.80–9.22; p 0.001) and 0.63 points (95% CI 0.50–0.75; p 0.001) from baseline, respectively (Figure 3, eTable 3). For the 16 subjects from this subset, who also had data at 3 years, the average FMA-UE and WMFT-FAS scores improved by 7.47 points (4.84–10.11; p 0.001) and 0.69 points (0.50–0.89; p < 0.001) from baseline, respectively. Age, sex, time poststroke, and paretic side were not significantly associated with change in FMA-UE or WMFT (eTable 3). FMA-UE and WMFT improvement remained stable between years 1 and 2 (Figure 3). FMA-UE and WMFT changes between groups for this subset during the blinded phase are presented in eTable 4. Thirty-seven of 49 participants (76%) demonstrated a clinically meaningful improvement in one or both the FMA-UE and WMFT at 2 years (eFigure 1). There were significant improvements from baseline in 5 of 7 measures of participation and quality of life (SIS-ADL, SIS-Hand, MAL-AOU, MAL-QOM, SS-QOL) at 2 years (eAppendix 1, eTable 5, eFigures 2 and 3).

Figure 3. Long-Term Impairment and Activity Changes in the Upper Extremity From Baseline.

Analyses limited to the 49 people who completed 2-year assessment. Fugl-Meyer Assessment-Upper Extremity (FMA-UE) (top) and Wolf Motor Function Test (WMFT) (bottom) change from baseline. Improvements from baseline remained stable from 1 to 3 years for the 38/49 participants who also completed the 1-year assessment and for the 16/49 participants who also completed the 3-year assessment (Note: all 16 individuals present for the 2- and 3-year assessment were also present at 1-year and are accounted for within the 38 individuals mentioned above). An additional 11 participants were exclusively assessed at 2 years for a total of 49 individuals (38 participants and 11 participants, respectively) included in this analysis. People randomized to active VNS (red) and control (blue) completed 6 weeks of in-clinic paired VNS + rehabilitation (for active VNS group, first shaded area; for controls, following crossover in second shaded area). Data were pooled for all participants from 1 year on (green). *Data previously reported.^5,6 VNS = Vagus nerve stimulation.

Discussion

These findings demonstrate that paired VNS therapy provides sustained long-term improvements in UE motor function among patients with moderate-to-severe chronic ischemic stroke. The long-lasting, clinically meaningful gains observed in UE motor impairment and function, along with improvements in participation and quality of life, support the clinical relevance of the benefits of Paired VNS therapy. Notably, while many interventions for chronic stroke may yield transient gains, durable changes that persist over time reflect a fundamental shift in the nervous system's function, a feature that is a priority for therapeutic development and clinical practice. Continued out-of-clinic engagement with self-activated VNS therapy may be a key contributor to the observed sustained benefits and is a noteworthy component of this intervention by providing an avenue for individual accountability and motivation. Performance of context-specific therapeutic tasks in the home and community may initially be facilitated by the therapy program but, over time, may transition to the naturalistic resumption of valued daily activities, thereby contributing to sustained improvements over several years.

Our 2-year results reinforce the value of paired VNS therapy. Over three-fourths of participants achieved clinically meaningful UE gains at 2 years. Notably, a subset of individuals demonstrated clinically meaningful improvement between years 1 and 2, including participants who either attained further MCID-level gains during this time or met MCID criteria for the first time at 2 years, having not achieved it by 1 year. These findings suggest not only durable retention of treatment effects but also the potential for continued recovery in some individuals who remain engaged in paired VNS tasks over time.

Limitations of this study are the lack of a control comparator and the unblinded nature of the assessments during the long-term phase. Minimal data are available to report the nature or frequency of participants' engagement in self-activated therapy during the second year, although such details would aid the interpretation of intersubject differences in long-term retention of functional gains.

Appendix. Coinvestigators

Name	Location	Role	Contribution
Rushna Ali	Mayo Clinic, Rochester, MN, the United States	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Benjamin L. Brown	University of Sheffield and Sheffield Teaching Hospitals, the United KingdomOchsner Neuroscience Institute, Covington, LA, the United States	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Louis DeMark	Brooks Rehabilitation, Jacksonville, FL, the United States	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Anand Dixit	The Newcastle Upon Tyne Hospitals NHS Trust, the United Kingdom	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Wayne Feng	Duke University, Durham, NC, the United States	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Gerard E. Francisco	University of Texas at Houston McGovern Medical School, TIRR Memorial Hermann, Houston, TX, US	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Leigh R. Hochberg	Massachusetts General Hospital, Harvard Medical School, Boston, MA, the United States	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Steven A. Kautz	VA Health Care System, Medical University of South Carolina, Charleston, the United States	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Charles Y. Liu	University of Southern California, Keck School of Medicine	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Arshad Majid	University of Sheffield and Sheffield Teaching Hospitals, the United Kingdom	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Michael W O'Dell	Weill Cornell Medicine and NeuroRehabilitation Consultants, NY, the United States	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Jessica Redgrave	Sheffield Teaching Hospitals NHS Trust, the United Kingdom	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB
Duncan L. Turner	University of East London, the United Kingdom	Site investigator	Study support, contributing participants, general supervision of study personnel, and data collection in VNS REHAB

Author Contributions

T.J. Kimberley: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. I. Vora: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. S.C. Cramer: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. S.L. Wolf: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. J. Dawson: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.

Study Funding

The VNS-REHAB pivotal trial was funded by MicroTransponder Inc.

Disclosure

T.J. Kimberley serves as a consultant for NeuroTrauma Sciences, Iota Biosciences, Myrobalan, MicroTransponder, and Mental Health Ritual; I. Vora serves as a consultant for MicroTransponder, Inc.; S.C. Cramer serves as a consultant for Alevian, Astellas, Bayer, BlueRock Therapeutics, BrainQ, Constant Therapeutics, Medtronic, MicroTransponder, Myomo, Myrobalan, NeuroTrauma Sciences, Simcere, and TRCare; S.L. Wolf serves as a consultant to SAEBO, Inc.and MicroTransponder; J. Dawson reports no disclosures. Go to Neurology.org/N for full disclosures.