Device and Procedure for Closed-Loop Endovascular Vagus Nerve Stimulation and Evoked Response Sensing

Abstract

BACKGROUND AND OBJECTIVES:

Vagus nerve stimulation (VNS) is a United States Food and Drug Administration–approved treatment for refractory epilepsy, stroke rehabilitation, chronic or recurrent depression as adjunctive therapy, and rheumatoid arthritis. It is also being explored as a potential treatment modality for heart failure and relapsing-remitting multiple sclerosis among other disease states. The current mode of stimulation delivery employed widely is the use of nerve cuffs mounted directly on the nerve via an invasive cervical cutdown procedure. Furthermore, the feedback from the nerve upon stimulation delivery is not monitored. It has also been reported that the internal fascicular organization along the vagus nerve is highly variable and therefore different groups of fascicles can be selectively activated based on electrode position. However, it is not possible to change the position of the nerve cuff without additional surgery if placed in a suboptimal location.

METHODS:

In this work, we propose the development of a device for closed-loop endovascular VNS that helps mitigate the challenges highlighted earlier and demonstrate this in a preclinical model on 4 subjects. In this minimally invasive procedure, a resheathable stent form factor is deployed within the jugular vein, providing an opportunity for intraprocedural trialing.

RESULTS:

We demonstrate accurate detection of nerve activation via endovascular evoked compound action potential along with an indication of the specific fibers being activated. Furthermore, this procedure also offers an opportunity to reduce side effects including muscle activation that is seen in current VNS procedures by tuning the applied stimulation.

CONCLUSION:

This closed-loop endovascular VNS stimulation platform technology opens up a new paradigm for ongoing research in peripheral nerve stimulation, affecting several disease states.

ABBREVIATIONS:

CLEVNS

closed-loop endovascular VNS

ECAP

evoked compound action potential

EJV

external jugular vein

FDA

US Food and Drug Administration

IJV

internal jugular vein

INS

neurostimulation system

PEEK

polyetheretherketone

VN

vagus nerve

VNS

vagus nerve stimulation.

Many neuromodulation therapies target different structures in the brain either invasively through deep brain stimulation^1,2 or noninvasively through transcranial stimulation transcranial direct current stimulation, transcranial alternating current stimulation, and transcranial magnetic stimulation.³ However, many research groups are interested in vagus nerve stimulation's (VNS) neuromodulation potential, investigating its mechanisms through projections to the brain (afferent fibers) and projections to the periphery (efferent fibers).^4-6 VNS has been a prominent therapy in peripheral neuromodulation⁷; however, the surgical process is invasive, which results in higher costs for the hospital, need for experienced surgeons, and patient agreement to an involved procedure.^8,9 Historically, VNS has been delivered through a helical cuff electrode implanted through a cutdown surgery in the cervical vertebral region.¹⁰ This therapy was originally developed by Cyberonics and later acquired by LivaNova.¹¹ The therapy is US Food and Drug Administration (FDA) approved for the treatment of refractory epilepsy (1997)¹² and the treatment of chronic or recurrent depression as adjunctive therapy (2005).¹³ Both of these indications are believed to modulate afferent pathways to the nucleus tractus solitarius.^14,15 More recently, CVRx was awarded FDA approval (2019) for their baroreflex therapy, which stimulates the vagus nerve (VN) baroreceptors for refractory heart failure.¹⁶ This therapy is implanted more cranial than the LivaNova VNS helical electrodes and modulates efferent projections to the heart. Multiple research groups are working to uncover the extensive topography of the VN, as variation within human and across animal models differs.^17-22 Others have made efforts to better characterize the adverse effects such as hoarseness and coughing that many experience after VNS implantation and therapy activation.^23,24 Although over 125 000 people have received LivaNova's VNS therapy,²⁵ the invasive nature of the procedure, including the cosmesis issues, could deter certain patient populations. Nicolai et al²⁶ demonstrated catheter-based endovascular VNS via the internal jugular vein (IJV) using an Orion cardiac mapping catheter in a swine/porcine model. They successfully stimulated the VN, and recorded compound action potentials (using intrafascicular electrodes placed into the VN) and motor evoked potentials of neck muscle (using bipolar needle electrodes) in an acute intraoperative experiment. Other studies^27-30 have described the benefits of VN fiber recruitment and closed-loop systems using conventional cuff electrodes. In a small cohort clinical study, Hays et al³¹ also demonstrated the benefits of closed-loop VNS to improve upper-limb recovery in individuals with chronic stroke.

Here, we propose an endovascular approach to stimulation of the VN through electrodes mounted on a resheathable stent. Conventional VNS therapies rely on direct contact with the VN. We have developed a novel 16 electrode stent lead, which allows for minimally invasive placement under fluoroscopic guidance through subclavian access that can be used to sense evoked compound action potential (ECAP) endovascularly. These ECAP recordings create a potential to provide closed-loop endovascular system to deliver VNS therapy. We aim to demonstrate that this procedure is feasible and effective compared with traditional cuff stimulation on multiple subjects using an acute ovine model.

METHODS

Preclinical Studies in Ovine Model

Ethical Statement

This work was conducted at the Physiological Research Laboratory (Medtronic, PLC), which is accredited by the association for Assessment and Accreditation of Laboratory Animal Care. The study protocol was reviewed and approved by the Institutional Animal Care and Use Committee. Note that no other IRB/ethics committee approval or patient consent was needed for this study.

Animal Housing and Care

Animals were housed and received humane care in accordance with the Guide for Care and Use of Laboratory Animals published by the National Institute of Health (National Institutes of Health Publication 85-23, revised 1996). Animals underwent a 10-day quarantine and acclimation period, physical examination, and clinical pathology evaluation to confirm normal health status before surgical interventions.

Generalizability/Translation

An ovine model was chosen to evaluate the feasibility of closed-loop endovascular VNS (CLEVNS). Although the ovine model lacks the IJV, the adult ovine external jugular vein (EJV) has a comparable diameter range and vasodilatory capacity to the adult human IJV.³² Furthermore, the ovine EJV does not reside within the carotid sheath, like a human IJV, and accordingly has greater mobility with respect to the VN than the human IJV, and therefore can serve as a conservative acute and chronic model as this concept is refined. Several acute and chronic preclinical studies have used the ovine model in deep brain stimulation where long-term survival of these animals were desired.^33-35 Deployment of the CLEVNS system in the ovine model is shown in Figure 1. Heparin sodium was administered up to 150 units/kg. Initial dose was given to target the intended activated clotting time, and the bolus was repeated approximately every 30 minutes as needed.

FIGURE 1.

Ovine CLEVNS system deployment. The deployment of the CLEVNS stent within the jugular vein in close proximity to the cervical vagus nerve in an ovine model is shown. CLEVNS, closed-loop endovascular VNS; CN, cranial nerve; INS, implantable neurostimulation system.

Protocol Registration

A detailed protocol, including research questions, study design, and analysis plan, was prepared before the study, reviewed, and approved by the Institutional Animal Care and Use Committee at the Physiological Research Laboratory (Medtronic, PLC).

Data Access

Data from this study are housed within Medtronic, PLC, and data access can be furnished upon request.

Fabrication of CLEVNS System

The CLEVNS system consists of 5 key components to deliver the therapy in the IJV. The components include (1) resheathable stent to accommodate 16 electrodes, (2) directional electrodes to steer the stimulation toward the nerve region, (3) leads that transmit the stimulation to the electrodes, (4) Percept implantable neurostimulation system to deliver the stimulation, and (5) delivery system that includes a 12 F sheath and a 10 F delivery catheter through which the above components are delivered to the target vessel as shown in Figure 2.

FIGURE 2.

Device fabrication. Various components of the CLEVNS system, including the stent, delivery system, leads, and electrodes, are shown. CLEVNS, closed-loop endovascular VNS.

The resheathable stent is a cylindrical structure constructed from nickel-titanium (nitinol) alloy with sufficient oversizing relative to the target vessel diameter to achieve good wall apposition. The closed cell structure in the stent is designed to accommodate 16 radiopaque electrodes that are mounted on projections from the inside of the cell and insulated from the stent body. The electrodes are mounted at 5-8 mm distances from each other in both circumferential and axial directions to allow distinct electrode pairs to function as stim and sense electrode pairs. The stent includes a radiopaque distal antenna feature that indicates circumferential midline of the 4 × 4 electrode grid in the stent. The end user can use the radiopaque distal antenna to rotationally center the electrode grid, under fluoroscopy, closest to the nerve per preprocedural planning.

One half of each platinum-iridium directional electrode is laser-textured to increase the charge injection capacity and the other half is coated with an insulating material, as shown in Figure 2. This is to direct the stimulation field away from the vessel lumen, toward the nerve.

Fabrication of the lead system consists of 16 small, well-insulated, conductor wires that are joined to connector rings on the proximal end and to the electrodes on the distal end. The conductor wires are welded to the connector rings and electrodes. The conductor pathway for each of the 16 circuits is contained in tubing that maintains insulation from the exterior of the lead system to enable precise control of delivering stimulation pulses and receiving sensing signals. The main lead body is constructed with polyurethane tubing from the proximal end to the distal end. The connector rings with welded circuit wires are insert-molded using micromolding techniques to fully encapsulate them and create an isodiametric proximal end with a defined spacing between the rings that will fit the connection system of the neurostimulator device.

The delivery system consists of a 10 F delivery catheter placed through a 12 F introducer sheath. The delivery system also consists of a multilumen polyetheretherketone pusher that helps with deployment and resheathing of the CLEVNS system.

Preprocedural Planning and Access

An inherent challenge with the using ovine EJV model for endovascular VNS is that it does not reside within a well-formed carotid sheath. Due to this, there is significant variation in the VN to EJV distance with neck motion. To mitigate this, a preoperative MRI procedure was implemented wherein the cervical anatomy of the sheep was scanned to obtain relative nerve to EJV distance and determine optimal location (both axial and rotational) for placement of the CLEVNS device in reference to the cervical spine landmarks, as shown in Figure 3A. During fluoroscopic procedure, this cervical spine landmark reference helped register learnings from the MRI images. During ovine acute studies, access was obtained to the EJV using a 12 Fr introducer sheath. A 0.035″ guidewire was introduced through the sheath and the delivery catheter containing the CLEVNS device was tracked along the guidewire up to the optimal cervical level identified during preprocedural planning. Once the electrode grid is axially centered adjacent to this identified location, the device is partially deployed until the antenna (shown in Figure 2) is exposed. The antenna is rotationally aligned in the direction of the VN, and the CLEVNS stent is fully deployed as shown in Figure 3B. The apposition of the stent to the vessel wall is confirmed using a contrast run.

FIGURE 3.

MRI procedural planning and access. A, Axial and coronal views of the cervical neck (ovine). Once the ideal (shortest) nerve to vessel distance is identified on the axial view, the cervical level is confirmed on the coronal view. B, The ceremony of use for the deployment of the CLEVNS stent via the jugular vein in the ovine model. CLEVNS, closed-loop endovascular VNS.

Intraprocedural Trailing

The CLEVNS device consists of 16 electrodes distributed along the stent in a 4 × 4 matrix. Once the device is deployed at the intended cervical level with desired axial position and circumferential orientation based on preprocedural planning, an ECAP mapping process is used wherein stimulation is applied on one electrode pair and propagation of resulting ECAP signal is monitored along adjacent pairs, as shown in Figure 4A. This helps identify optimal electrode pairs with the shortest distance between the nerve and vessel. To confirm that the sensed ECAP signal is neural and not muscle activity, a neuromuscular blocker (atracurium) was administered. The ECAP signal was compared before and after administration, as shown in Figure 4A.

FIGURE 4.

Optimal stimulation and sensing configuration during implantation. A, Device orientation during implantation within the vessel is determined through observation of ECAPs under a stimulation/sensing ECAP mapping process. Atracurium (a neuromuscular paralytic) is delivered to suppress EMG activity that confounds with vagus nerve ECAPs. Optimal bipolar pairs sense the largest ECAP and demonstrate greatest proximity to the vessel. Columnwise adjacent optimal sensing electrode bipolar pairs may demonstrate a consistent propagation delay of the ECAP as a function of distance, indicating the device is parallel to the vagus nerve. Within the map, the spatial coverage and bipolar electrode pairs of an 8e 4 × 2 electrode layout are limited compared with a 16e 4 × 4 electrode layout. B, Vessel coverage difference between 2 × 4 and 4 × 4 electrode maps on a CLEVNS stent. Note that this illustration is for an 18 mm stent in a 15 mm vessel. ECAP, evoked compound action potential; EMG, electromyograph; INS, implantable neurostimulator.

RESULTS

The neural component of the signal is retained, thereby indicating activation of the VN bundle. Furthermore, benefit of a 4 × 4 or 16-electrode configuration over 4 × 2 or 8-electrode configuration is shown in Figure 4B. The field distribution is increased significantly using a 16-electrode configuration, thereby reducing the deployment/resheathing cycles and improving the ability to stimulate the nerve more effectively without multiple ECAP mapping processes.

Endovascular ECAP Across Subjects and Signal Propagation

Endovascular ECAP obtained across multiple ovine subjects (n = 4) is shown in Figure 5. Figure 5A shows sensed ECAP signals across multiple ovine subjects with varying current amplitude range (∼0.1-5 mA). The optimal stim/sense electrode pair was determined as described in the previous section and these pairs were used to characterize ECAP response with current amplitude ramp. Figure 5B shows the sensed ECAP peak comparison between subjects. This variation can be attributed to the varying nerve to vessel distance in the ovine model due to the EJV not residing in the carotid sheath. We do not anticipate this in human subjects where the carotid artery, IJV, and VN are well contained within the carotid sheath. The proximity between the VN and IJV is reported between 0 and 2 mm in literature.³⁶ Figure 5C demonstrates ECAP signal propagation across 4 ovine subjects. Note that interelectrode distance and ECAP signal propagation can also be used to identify signal morphology related to muscle response. The neuromuscular blocker (atracurium) was only for verification in early work. Propagation delay as a function of interelectrode distance is also shown. Linear regression of the mean propagation delay approximates an ECAP propagation velocity of 3.85 m/s. Based on previous literature,³⁷ this is consistent while activating B fiber type of the VN bundle where conduction velocity is between 3 and 14 m/s.

FIGURE 5.

Endovascular ECAP across subjects and signal propagation. A, ECAP growth curves across 4 acute ovine subjects. Each subject's ECAPs are captured using adjacent bipolar stimulation and sensing electrode pairs as shown in the 4 × 4 device legend. Significant increase in ECAP observed between 0.1 mA and higher amplitudes (*P < .003, t-test). B, ECAPs are shown across 4 subjects with similar propagation delays. Statistically significant correlations in all subjects between stimulation amplitude and ECAP peak amplitude (*P < .0011, Pearson correlation). C, Propagation delay measured across multiple electrode bipolar pairs within each subject. Linear regression of the mean propagation delay for each electrode pair distance is shown. The slope of the linear regression approximates the ECAP propagation speed of 3.85 m/s. ECAP, evoked compound action potential.

DISCUSSION

This study demonstrates the features and advantages of a CLEVNS system using an optimized stent-based platform. VNS is an FDA-approved therapy for epilepsy, stroke rehabilitation, depression, and rheumatoid arthritis, with other applications being explored. Therefore, there is tremendous interest in expanding access to care and increasing patient adoption for these therapies. With the perceived invasiveness of current nerve cuff–based VNS therapy and its inherent complications, including scarring and gliosis, adoption for this approach has been limited. Figure 6A shows the current standard of care for VNS using a nerve cuff. An endovascular procedure as proposed in this work and shown in Figure 6B demonstrates feasibility of a minimally invasive catheterization procedure for a closed-loop, targeted delivery of stimulation to the cervical VN. An inherent benefit of using a stent-based form factor for endovascular stimulation is the ability of the stent to endothelialize with time, which can provide better, more stable stim/sense capabilities, and demonstrate an improved safety profile. Chronic ECAP stability and effects of endothelialization are planned to be studies in future chronic studies using an implantable form factor. This novel technology not only delivers targeted stimulation but also has the ability to tune the therapy based on sensed ECAPS, thereby providing an opportunity to deliver personalized care. Furthermore, when accessed via the subclavian vein, which is a commonly used access point in several interventions including placement of cardiac pacing leads, it eliminates the need for tunneling, thereby reducing procedural time and burden. We also demonstrated activation of specific nerve fibers within the VN bundle as confirmed through the endovascularly sensed ECAP propagation speed. The use of specific stimulation parameters and/or directional electrodes provides an opportunity to further study activation of specific nerve fibers, which may provide reduced side effects for patients. Preprocedural planning along with use of 4 × 4 16-electrode design provides effective field distribution to target the nerve effectively. The CLEVNS system has the potential to open a new paradigm in peripheral endovascular stimulation.

FIGURE 6.

CLEVNS system for endovascular vagus nerve stimulation. A, Current standard of care for vagus nerve stimulation using nerve cuffs is shown. The procedure requires a cervical cut down, dissection of the carotid sheath, and placement of the cuff on the nerve. After the placement of cuffs, the lead body is tunneled to the INS pocket. B, Proposed CLEVNS system is shown where the endovascular system is placed using a minimally invasive catheterization procedure. No tunneling is needed since the exit of the leads from the subclavian vein is adjacent to the INS pocket. CLEVNS, closed-loop endovascular VNS; CN, cranial nerve; INS, implantable neurostimulation system.

Limitations

This work highlights the feasibility of CLEVNS via ECAP sensing. However, chronic and clinical evaluations are necessary to study the effects of this approach for translational therapeutic benefits. In a 14-year single-center experience reported by Van Schootan et al,³⁸ surgical complications for conventional cuff-based VNS included infection, vocal cord paresis, hematoma, and hardware failure. They reported complication rate after primary implantation at 13.4%, 2.5% for pulse generator replacement, 21.4% for lead revision, and 27.3% for complete VNS removal. However, venous stenting has complication concerns, including IJV cannulation, stent migration, stent fracture, in-stent stenosis/compression, infection, acute thrombosis, need for revision/explant procedures, and antithrombotic therapy in other applications.^39,40 These need to be investigated further in future chronic studies. Furthermore, although a small cohort clinical study by Hays et al³¹ demonstrated the benefits of closed-loop VNS to improve upper-limb recovery in individuals with chronic stroke, future large cohort studies are needed to demonstrate improvement in patient outcomes with ECAP-guided closed-loop VNS.

CONCLUSION

This work demonstrates the ability of the CLEVNS system for peripheral nerve stimulation, specifically for cervical VNS. The system is optimized to stimulate the nerve while obtaining ECAPs to tune the therapy delivered. The ability of this system to target specific nerve fibers offers the opportunity for reduced side effects. However, further work is needed for chronic evaluation to demonstrate long-term stability of this system, which is critical for VNS therapy.

Funding

All work was funded by Medtronic, PLC. Authors received no specific research grant from external funding agencies.

Disclosures

All authors except AAK are employees and stockholders of Medtronic PLC. AAK consults with Medtronic.

COMMENTS

The investigators have made significant progress toward an endovascular vagal nerve stimulator (VNS). The present paper provides compelling evidence that such a device is practical, deploying it in the jugular vein of sheep and showing that they can stimulate the nerve by recording from the nerve with the same device. The “closed loop” nature of the device, referring to its capacity to both stimulate and record, is critical to positioning the electrode array in apposition to the nerve. That is, confirmation that stimulation is creating a propagating nerve action potential confirms that the stent is rotated such that the electrodes bracket the nerve position. While technically meeting the definition of “closed loop,” the device does not have the capacity to adjust its output in response to alterations in the target disease, as the term is usually used in the field of neuromodulation.

Open placement of VNS is a safe and routine procedure, begging questions about the market for an endovascular vagal stentrode. There is certainly a cosmetic advantage, avoiding visible scarring in the neck. Given the low complication rates associated with the current design, it remains to be seen whether the availability of an endovascular system will drive demand and, as the authors suggest, expand the impact of VNS therapies by attracting more patients. One area that would provide obvious benefit is lead revisions. While VNS implantation is relatively easy the first time, removal and replacement particularly after an infection is far more challenging. The authors tout the ability of the closed loop approach to selectively stimulate specific subpopulations of fibers. However, it would seem that preferential stimulation of specific axons could be achieved by the current systems once the appropriate waveforms are defined. Moreover, it is not clear whether this capacity will impact the efficacy of VNS.

The current electrode design provided by the different companies is surprisingly similar. In fact, the intellectual property seems to be based on the companies' claims for treatment of individual disorders: depression, epilepsy, stroke rehabilitation, and arthritis. To this extent, it will be interesting to see how the adoption of this approach differs between the various companies/applications.

Nicholas Boulis

Atlanta, Georgia, USA