Vibrotactile Stimulation in Lower Extremity Amputation Patients Using a Noninvasive Vibration Device: A Pilot Study

Abstract

Background:

Advances in surgery and prosthetic technology have improved limb control for amputation patients, but restoring sensory and proprioceptive feedback remains a challenge. This study evaluated the application of a vibrotactile sensory feedback device that aims at enhancing gait and reducing phantom limb pain (PLP) in lower extremity (LE) amputation patients who have undergone targeted muscle reinnervation (TMR).

Methods:

Four male LE amputation patients, 25–68 years of age, who underwent primary TMR, were fitted with a vibrotactile device using pressure sensors located on the sole of the prosthesis and a vibrating actuator on the proximal LE. This device incorporates vibrational stimuli when walking in a prosthesis for real-time sensory and proprioceptive feedback. Participants used the device alongside their regular prosthesis for 31 days. Pain, function, mental health, and satisfaction metrics were assessed using the visual analog scale, patient-reported outcomes measurement information system surveys, and various gait tests at baseline and follow-up. Baseline demographic, surgery, and comorbidity data were collected from chart review.

Results:

Three of 4 patients reported reduction in PLP and improvements in gait and device satisfaction. One young patient, who was highly active, showed limited improvement compared with the others. One patient experienced a reduction in anxiety and depression.

Conclusions:

The vibrotactile feedback device demonstrated potential in improving PLP and gait among LE amputation patients who underwent primary TMR. Patient activity levels and psychological factors likely play important roles in the clinical effectiveness of the device. Future studies should focus on personalizing interventions based on patient profiles and exploring long-term benefits.

Takeaways

Question: Can a noninvasive vibrotactile sensory feedback device improve gait and reduce phantom limb pain in lower extremity amputation patients who underwent primary targeted muscle reinnervation (TMR)?

Findings: In this pilot study of 4 male lower extremity primary TMR amputation patients, 3 reported reduced phantom limb pain and improved gait after using a vibrotactile feedback device for 31 days. Moderately active patients showed the most benefit.

Meaning: Vibrotactile feedback devices may aid in phantom limb pain and gait in primary TMR amputation patients, particularly those with moderate activity levels. Personalized approaches considering individual factors may optimize outcomes.

INTRODUCTION

In the United States, approximately 150,000 patients undergo a lower extremity (LE) amputation annually.¹ LE amputation patients commonly experience neuropathic pain, including phantom limb pain (PLP) and residual limb pain, along with decreased mobility and altered mental well-being, affecting multiple facets of patients’ lives.^2–7 During the past few decades, advances in prosthetic devices have contributed to significant improvements in the surgical care of amputation patients. Innovations such as osseointegration⁸ and advanced nerve surgical techniques, including targeted muscle reinnervation (TMR) and regenerative peripheral nerve interfaces (RPNIs), have enhanced the function of LE amputation patients.^9–13

Despite these technological advances in control of artificial limbs, restoring sensory and proprioceptive feedback remains a challenging aspect of achieving balanced prosthetic control. A recent development in this field is targeted sensory reinnervation (TSR), which is a nerve transfer technique that involves reinnervating a specific skin area with another sensory nerve.¹⁴ Additionally, the application of a vibrotactile feedback system connected to the LE prosthetic socket has shown promising results in restoring proprioception, leading to improved walking ability and reduced PLP.¹⁵ However, TSR is a novel surgical technique that may not always be feasible and is not accessible to all individuals with LE amputations. Additionally, it remains unclear whether TSR is essential to fully realize the benefits of this device. Therefore, the primary aim of this pilot study is to investigate whether this vibrotactile sensory feedback device can improve gait and reduce PLP in a pilot group of major LE amputation patients who underwent primary TMR, but not TSR, at the time of amputation. Additionally, we aimed to assess the satisfaction and impact of the device on patients’ well-being.

MATERIALS AND METHODS

Patient Population

Following institutional review board approval (2023P001047), we contacted eligible adult patients with unilateral transtibial or transfemoral amputation who underwent TMR at amputation at our peripheral nerve clinic.¹⁶ Inclusion criteria required current care at a local prosthetic clinic, some degree of PLP, and community ambulation capability (K-2 or higher). (See table, Supplemental Digital Content 1, which displays the lower limb extremity prosthesis Medicare Functional Classification Levels [K-levels], http://links.lww.com/PRSGO/E41.) No restrictions were placed on pain duration, intensity, or the indication for amputation. The study trial period occurred from February to March 2024. A maximum of 8 devices were provided for this study. Eleven patients were contacted by telephone, and 4 agreed to participate.

Device

The noninvasive vibrotactile sensory feedback device consists of a foot sole with integrated sensors and transmitters, and a cuff or pant with 4 vibration motors placed on the patient’s LE residual limb (Fig. 1). The foot sole contains 4 sensors—3 under the forefoot and 1 on the heel—corresponding to the vibration motors. During walking, sensors transmit pressure signals to the motors, providing synchronized feedback that allows perception of ground contact without visual confirmation. An external manufacturer provided the devices; our team had no role in their development.

Fig. 1.

Overview of study procedure. At baseline (day 1), patients underwent assessment 1 (PROMs + gait tests). After this, the device was fitted. Patients were instructed to use the device for at least 10 hours daily, in addition to their prosthesis, and to log their pain levels and daily device usage. At the follow-up (day 31), patients repeated the PROMs and gait tests while still using the device. PROMs, Patient-reported outcome measures

Study Procedure

Patients were invited for a baseline and a follow-up assessment at a prosthetic clinic (Fig. 1). At baseline, patients completed demographics surveys, patient-reported outcome measures, and gait tests with their regular prosthesis. Subsequently, patients received the device with fitting and usage guidelines from a manufacturer’s representative. Patients were instructed to use the device for at least 10 hours daily for 31 days with their regular prosthesis. During this period, patients were given a pain diary to log their daily PLP and hours of device use. After 31 days, patients returned to the prosthetic clinic for a follow-up visit, during which the surveys and gait test were again administered. The patients were instructed to use the device through the final day of follow-up.

Data Acquisition

All data were recorded using Research Electronic Data Capture (REDCap 12.4.27; Vanderbilt University) (HIPAA-compliant). Demographics (sex, age, race, education, relationship status), surgery (level and indication of amputation, surgical history), comorbidity data (body mass index, K-level, regular prosthetic wear time, Elixhauser Comorbidity Index, diabetes (type I or II), alcoholism, smoking, psychiatric comorbidities (post-traumatic stress disorder, depression, anxiety), and postoperative opioid and neuromodulator use were obtained through the demographic survey or obtained through chart review. The Medicare Functional Classification Level (or K-level) for LE prostheses is a rating system used to indicate a patient’s potential for different types of prosthetic devices and the associated reimbursement levels. Patients are classified into 5 categories based on activity level: K-0 (no ability to ambulate), K-1 (limited ability to ambulate), K-2 (moderate ability to ambulate), K-3 (good ability to ambulate), and K-4 (ability for high-demand ambulation) (Supplemental Digital Content 1, http://links.lww.com/PRSGO/E41).

At baseline, we administered the visual analog scale (VAS) for momentary pain, patient-reported outcomes measurement information system (PROMIS) pain interference short form (SF) 4a for pain’s impact on daily life, and PROMIS pain intensity SF3a for average pain over 7 days, all specific to PLP. For functional outcomes, the PROMIS physical function SF6b was used; for mental health, the PROMIS anxiety SF4a and PROMIS depression SF4a; for device satisfaction, the orthotics and prosthetics users survey satisfaction with device. The patient global impression of change (PGIC) was used to evaluate the patient’s perceived changes in both PLP and mobility. The PGIC scale ranges from −3 (indicating “very much declined”) to +3 (indicating “very much improved”), with 0 representing “no change.” This measure was only administered at follow-up. The following gait tests were conducted at both baseline and follow-up: the timed up-and-go test (TUGT), the 10-meter walk test (MWT-10), the four-square step test (FSST), and the 2-minute walk test (MWT-2). (See table, Supplemental Digital Content 2, for an overview of surveys and tests conducted, http://links.lww.com/PRSGO/E42.)

Statistical Analysis

For data analysis, Stata IC (version 16, StataCorp LLC, College Station, TX) was used. All numeric data are presented with mean and SD, and categorical variables display raw numbers and percentages per category. No comparative analysis was conducted due to the limited sample size. All analyses were conducted in adherence to the SAMPL guidelines.¹⁷

RESULTS

The 4 included patients were men, with a mean age at the time of amputation of 52.5 ± 19.3 years (range: 25.8–69.0 y). All 4 patients sustained an injury that led to chronic pain or limited functionality, ultimately resulting in an elective amputation. The mean time between initial injury and LE amputation with primary TMR was 25.5 ± 27.7 years (range: 1.7–53.8 y). The mean time between amputation and primary TMR to final follow-up was 1.1 ± 0.5 years (range: 0.4–1.6 y). Activity levels at the time of the study trial ranged from K-2 (patient 2) to K-4 (patient 3). The average Elixhauser Comorbidity Index was 4.7 ± 2.2, and no opioids or neuropathic analgesics were used at the time of study evaluation. All patient demographics, comorbidity, and surgery characteristics can be found in Table 1.

Table 1.

Demographics, Surgery, and Comorbidity Characteristics

Patient 1

A 52-year-old man sustained initial injuries after a snowmobile accident, undergoing right knee arthroscopy, allograft anterior and posterior cruciate ligament (PCL) reconstruction, and open medial collateral ligament reconstruction with allograft. He presented with a chronic infection of his knee. Despite numerous attempted salvage procedures, the patient continued to experience chronic infection and developed osteomyelitis, pain, limited knee extension, and foot drop. This situation resulted in an elective right transfemoral amputation with direct TMR of the saphenous, tibial, and common peroneal nerves 1.7 years after his initial accident. His final follow-up was 1.4 years postamputation with TMR.

Patient 2

A 64-year-old male fire inspector presented with a long history of right ankle pain dating back to a childhood bicycle injury. Despite numerous procedures, including arthroscopic debridement, microfracture, and total ankle arthroplasty, he continued to experience severe neuropathic foot and ankle pain. Peroneal tendon exploration, debridement, and tarsal tunnel release provided no relief. He underwent an elective right transtibial amputation including TMR of the saphenous, tibial, distal peroneal, and superficial peroneal nerves and RPNI of the sural nerve 53.8 years after the initial accident. His final follow-up was 5 months postamputation with TMR.

Patient 3

A 26-year-old man presented with a left ankle traumatic injury following a motor vehicle accident (MVA). The patient reported primary pain in the medial and lateral aspects of his left ankle during dorsi- and plantar flexion, as well as hypersensitivity in his toes and left arch, limiting his walking ability. Diagnostic local anesthetic blocks and lumbar injections provided no relief. He underwent an elective left transtibial Ertl amputation and direct TMR with nerve transfer of the saphenous, tibial, distal peroneal, superficial peroneal, and sural nerves 1.7 years after the MVA. His final follow-up was 1.0 years postamputation with TMR.

Patient 4

A 70-year-old man presented with chronic right foot and ankle pain following an open tibia/fibula fracture sustained during a previous MVA. Despite extensive operations for attempted limb salvage, including bone grafts, muscle flaps, skin grafts, and ankle fusion, his pain was debilitating. Conservative treatments, such as bracing, physical therapy, and injections did not provide benefit. Finally, 44.6 years after the MVA, he underwent an elective right transtibial amputation with TMR of the superficial peroneal, tibial, and sural nerves as well as an RPNI of the saphenous nerve. His final follow-up was 1.6 years postamputation with TMR.

Device Usage and Survey and Test Outcomes

Patients received the device at baseline and were instructed to use it for at least 10 hours daily until day 31 of follow-up (Fig. 2). Patient 1 used the device for 13.5 ± 1.3 hours/d; patient 2 used the device for 11.9 ± 1.5 hours/d; patient 3 used the device for 4.5 ± 3.5 hours/d for the first 8 days before he stopped using the device at day 9; and patient 4 used the device for 9.6 ± 5.0 hours/d but reported difficulty with usage of the device for a total of 7 days.

Fig. 2.

Device use during the study trial period.

The 4 patients were evaluated at baseline and follow-up for pain, function, mental health, and satisfaction metrics and gait testing. Surveys and tests were conducted at baseline (prosthesis only) and at follow-up (prosthesis + device). Outcomes are found in Table 2. Three of 4 patients reported an improvement in their VAS PLP scores (from scores of 5, 3 and 2 at baseline, respectively, to 0 at follow-up, Fig. 3A). Patient 4 showed no difference in his VAS scores. PLP-pain intensity (Fig. 3B) improved in 3 patients, with reductions of 7.2 (patient 1), 17.9 (patient 2), and 9.9 (patient 4), whereas patient 3 reported no change (Δ = 0). PLP interference (Fig. 3C) improved in 2 patients, with reductions of 4.4 (patient 1) and 17.3 (patient 2), whereas patients 3 and 4 reported no change (Δ = 0). Physical function (Fig. 3D) improved in 2 patients, with increases of 1.8 (patient 1) and 6.1 (patient 2), and remained unchanged in patients 3 and 4. Gait tests (Fig. 4) indicated improvement in all patients except for patient 3, who had the highest scores on all tests at baseline and follow-up. Specifically, patient 1 showed improvements in MWT-10 (Δ = 0.79), MWT-2 (Δ = 5.5), FSST (Δ=-1.98), and TUGT (Δ=-1.16), whereas patient 2 showed even greater improvements: MWT-10 (Δ = −5.71), MWT-2 (Δ = 44), FSST (Δ = −10.66), and TUGT (Δ = −8.39). Patient 4 also showed significant improvements in gait tests: MWT-10 (Δ = −1.26), MWT-2 (Δ = 9), FSST (Δ = −1.50), and TUG (Δ = −0.11). The PGIC ratings (Fig. 5) for pain and mobility showed improvement in 3 patients, with only patient 3 reporting no change.

Table 2.

Survey and Test Outcomes

Fig. 3.

Pain and function, for prosthesis vs device use. A, Visual analog scale for PLP. B, PROMIS pain intensity for PLP. C, PROMIS pain interference scores for PLP. D, PROMIS physical function.

Fig. 4.

Gait tests, for prosthesis vs device use. A, Gait timed up-and-go test. B, Gait 10-meter walk test. C, Gait FSST. D, Gait 2-minute walk test.

Fig. 5.

Patient global impression of change, for pain and mobility.

All patients were satisfied with the prosthesis and device (Fig. 6). Patient 4, who had baseline anxiety (59.5) and depression (55.7) scores, reported improvements at follow-up, with anxiety decreasing to 53.7 (Δ = −5.8) and depression to 51.8 (Δ = −3.9) (Fig. 7). The 3 other patients had no reported anxiety/depression.

Fig. 6.

OPUS satisfaction of device, for prosthesis vs device use. OPUS, Orthotics and Prosthetics Users Survey.

Fig. 7.

Mental health outcomes, for prosthesis vs device use. A, PROMIS anxiety. B, PROMIS depression.

DISCUSSION

In this pilot study, 4 LE amputation patients with TMR participated in a 31-day vibrotactile device trial. Two patients met recommended usage (≥10 h daily) and reported improved PLP, mobility, and satisfaction. One patient with a history of depression reported improvements in depression and anxiety levels.

The differential response among patients to the vibrotactile feedback device may be linked to their baseline activity levels and device engagement. Patients with moderate activity levels (patients 1, 2, and 4) demonstrated significant reductions in PLP and improvements in gait, whereas patient 3, who was highly active and had better baseline scores, showed minimal change. This difference may also be linked to cortical reorganization following TMR and amputation recovery. Specifically, the efficacy of sensory feedback devices may be influenced by the extent to which cortical representation of the missing limb has been adapted or stabilized.¹⁵ Vibrotactile feedback can influence the cortical maps associated with the residual limb, a phenomenon supported by research on sensory prostheses and cortical plasticity.¹⁸ This device may provide sensory inputs that help remap the cortical representation of the missing limb, potentially reducing PLP by modulating the sensory feedback pathways and engaging areas of the brain associated with proprioception and limb perception.^19,20 Patient 2, who had the most recent amputation (5-mo follow-up), demonstrated the greatest improvement in pain and gait outcomes. The other patients had follow-ups ranging from 1.0 to 1.6 years. Although this might indicate that the baseline outcomes were affected by the timing of the amputation, it is worth investigating whether earlier use of the device could be advantageous, given the timing of cortical reorganization. However, other factors affecting the speed of cortical reorganization may play a role, such as time since amputation, age, and overall health status.^21,22

The baseline psychological status and patient mindset also plays an important role in the response to sensory feedback interventions. Patient 4 had significant anxiety and depression at baseline, and did not report notable improvements in PLP outcomes. This aligns with findings that psychological factors can influence pain perception and treatment outcomes.^23,24 However, patient 4’s mental health improvements may suggest that the device’s benefits extend beyond physical and mechanical enhancements, potentially contributing to overall patient satisfaction and quality of life through enhanced mobility and gait. The limited effect on PLP observed, along with patient 4’s inconsistent use of the device, may highlight the importance of patient engagement in optimizing the efficacy of this modality.^25–27

Patient characteristics significantly influenced device effectiveness. The baseline activity level emerged as crucial—moderately active patients showed significant improvements, whereas our highly active, younger participant experienced limited benefits, suggesting those with excellent baseline function have less room for improvement. Psychological factors also proved important, as 1 patient with baseline anxiety and depression showed improvements in both mental health and function, indicating a bidirectional relationship between psychological well-being and device benefits. Future studies should assess psychological readiness and activity levels to optimize patient selection.

This study’s findings suggest that the vibrotactile feedback device is often helpful but may not be universally effective for all LE amputation patients. The variability in patient responses highlights the need for personalized approaches. Tailoring device usage to individual profiles—considering baseline activity levels, functional status, and psychological readiness—may optimize outcomes. For instance, patients with low or moderate activity levels may benefit more due to ongoing cortical adaptation and therefore have a more pronounced need for additional sensory and proprioceptive feedback.¹⁹ Conversely, highly active patients or those with advanced cortical reorganization may derive fewer benefits, as their sensory pathways may already be more robustly integrated and well-adapted to their prosthetic use.²⁸ Future research should focus on identifying patient characteristics that predict a better device This could involve stratifying patients based on their activity levels, psychological assessments, and baseline functional capacities to determine who benefits most from such interventions. Additionally, exploring the integration of motivational and psychological support strategies alongside the device could enhance patient engagement and potentially improve overall outcomes. Personalized interventions considering these factors could more effectively manage PLP and improve prosthetic function in LE amputation patients.

Previous studies indicate that vibrotactile feedback devices improve postural stability and gait symmetry by enhancing proprioception.^29–32 For instance, patients who underwent TSR used a prosthesis with a pressure sensor embedded in the sole.¹⁵ This sensor’s signals were transmitted to an actuator on the reinnervated skin area, allowing sensations from the prosthetic sole to be perceived as if originating from the lost foot. Without TSR, although sensory feedback is still evoked, it is typically perceived as coming from a nonphysiological location (such as the thigh, in this study), rather than the foot. Even without TSR, we observed improvements in proprioception, likely due to increased confidence in limb use and a reduced reliance on visual feedback. These proprioceptive improvements suggest that the device provides a more intuitive sense of foot position, even when the sensation does not perfectly replicate that of the lost limb.

Recent research continues to expand our understanding of TMR and RPNI applications. Roubaud et al³³ demonstrated outcomes for pain control in oncologic amputation patients, whereas ElAbd et al³⁴ found improvements in pain and functional outcomes across diverse populations. Palafox^35,36 showed that combined TMR and RPNI can be effective for trauma-related amputations. Modern electrical stimulation methods for RPNIs show promise for restoring sensory feedback, complementing vibrotactile approaches.³⁷ Assessment tools have evolved including virtual measuring systems and the LACE+ Index algorithm for predicting outcomes.^38–40 Recently, it was demonstrated that preoperative pain sketches can help predict pain outcomes after secondary TMR, reflecting a broader trend toward incorporating objective measurements and predictive analytics into surgical practice.⁴¹ These developments suggest that TMR and RPNI can benefit various populations, but patient assessment and selection using modern tools may optimize outcomes.

This study is an early report of use of a vibrotactile stimulation device in LE amputation patients undergoing TMR. Although TMR can not only treat^42,43 but also prevent neuropathic pain for many patients,^44,45 with 53.7% of primary TMR patients achieving pain prophylaxis (VAS ≤ 3), and 26.7% achieving complete pain relief,⁴⁶ it is not universally effective.⁴⁷ In this study, 3 of 4 patients reported reductions in PLP and improvements in gait within just 1 month, perhaps demonstrating the potential of these devices as an adjunct to primary TMR. However, due to the small sample size and lack of a control group, we hope these findings may prompt further robust research studies.

CONCLUSIONS

In this pilot study, a vibrotactile feedback device showed promise in reducing PLP and improving gait among LE amputation patients who underwent primary TMR. Patient activity levels and psychological factors appeared to influence outcomes. Future research with larger sample sizes is needed to validate these findings and establish optimal patient selection criteria.

DISCLOSURES

Dr. Valerio is a consultant for AxoGen, Inc., Checkpoint Surgical, Inc., and Integra Lifesciences, Inc. Dr. Eberlin is a consultant for AxoGen, Inc., Checkpoint Surgical, Inc., Integra Lifesciences, Inc., Tissium, Tulavi Therapeutics, Inc., and Biocircuit. Mr. Graham is a representative of Next Step Bionics & Prosthetics, Inc. The other authors have no financial interest to declare in relation to the content of this article. This work was in part supported by the Jesse B. Jupiter/Wyss Medical Foundation Endowment.

ACKNOWLEDGMENT

The figures were created with Biorender.com.

Supplementary Material

gox-13-e6788-s001.pdf ^{(70.5KB, pdf)}

gox-13-e6788-s002.pdf ^{(80.2KB, pdf)}