Vagus nerve stimulator revision in pediatric epilepsy patients: a technical note and case series

Vagus nerve stimulation (VNS) is an adjunctive treatment in children with intractable epilepsy. When lead replacement becomes necessary, the old leads are often truncated and retained and new leads are implanted at a newly exposed segment of the nerve. Direct lead removal and replacement are infrequently described, with outcomes poorly characterized. We aimed to describe our experience with feasibility of VNS lead removal and replacement in pediatric patients. Retrospective review examined 14 patients, at a single, tertiary-care, children’s hospital, who underwent surgery to replace VNS leads, with complete removal of the existing lead from the vagus nerve and placement of a new lead on the same segment of the vagus nerve, via blunt and sharp dissection without use of electrocautery. Preoperative characteristics, stimulation parameters, and outcomes were collected. Mean age at initial VNS placement was 7.6 years (SD 3.5, range 4.5–13.4). Most common etiologies of epilepsy were genetic (5, 36%) and cryptogenic (4, 29%). Lead replacement was performed at a mean of 6.0 years (SD 3.8, range 2.1–11.7) following initial VNS placement. Reasons for revision included VNS lead breakage or malfunction. There were no perioperative complications, including surgical site infection, voice changes, dysphagia, or new deficits postoperatively. Stimulation parameters after replacement surgery at last follow-up were similar compared to preoperatively, with final stimulation parameters ranging from 0.25 mA higher to 1.5 mA lower to maintain baseline seizure control. The mean length of follow-up was 7.9 years (SD 3.5, range 3.1–13.7). Removal and replacement of VNS leads are feasible and can be safely performed in children. Further characterization of surgical technique, associated risk, impact on stimulation parameters, and long-term outcomes are needed to inform best practices in VNS revision.


Introduction
Epilepsy is a debilitating neurological disorder with an incidence of 41 to 187 per 100,000 children, comprising 1% of all children globally [1][2][3][4][5]. Greater than one-third of these children fail to attain seizure freedom with antiepileptic drugs (AEDs) [6][7][8][9][10]. Epilepsy surgery can result in improved outcomes in these patients. However, some children are not a candidate for resective, ablative, or disconnection epilepsy surgery or do not respond [11], leading to significant impacts on quality of life [12].
Neuromodulation with vagus nerve stimulation (VNS) serves as an adjunctive treatment for these children [13][14][15]. VNS consists of an internal pulse generator connected to a stimulating electrode lead housed in helical coils wrapped around the vagus nerve [16]. Surgery involves neck dissection down to the carotid sheath and placement of the lead with three coils around the vagus nerve, which is then tunneled to a subcutaneous or submuscular pocket on the chest wall [17]. VNS placement No part of this work has been previously published.

3
is routinely performed on the left side to minimize risk of arrythmia and cardiovascular sequelae, as the right vagus nerve innervates the sinoatrial node [18]. Although the precise mechanism by which VNS functions is unclear, VNS may utilize the primary afferent connection of the vagus nerve to the nucleus tractus solitarius and projections to the locus coeruleus to increase downstream release of norepinephrine [19][20][21]. VNS is associated with a seizure reduction of at least 50% in over half of patients [14]. Treatment by VNS also decreases seizure severity, recovery time, daytime drop attacks, and hospitalizations and improves quality of life [22][23][24].
Malfunction, infection, or intolerable side effects may necessitate removal or revision of VNS hardware [16,[25][26][27][28]. While the pulse generator is easily retrieved, the helical leads are infrequently removed [16]. Need for frequent MR imaging has also been encountered as a motivation for VNS removal. Traditional reasoning against removing VNS leads includes convenience, unfamiliarity with the procedure, and concern for vagus nerve injury [29][30][31]. However, retained VNS hardware may increase the risk of infection, while certain lengths of electrode and cabling may preclude subsequent magnetic resonance imaging (MRI) evaluation [26,32]. VNS removal has been described in the adult literature, with studies demonstrating feasibility of complete VNS lead removal using blunt and sharp dissection or blunt dissection with electrocautery [26,30,31,33,34], while others demonstrated continued VNS efficacy following revision to replace leads [26,31,36,37]. We aimed to describe our experience with removal and same-site replacement of VNS leads in pediatric patients to further characterize the optimal surgical technique for VNS lead removal in the children.

Study design
A retrospective chart review was performed for all pediatric patients who underwent surgery to replace or revise VNS leads at a tertiary children's hospital from 2010 to 2018, allowing for at least two years of follow-up, through 2020. Patients were identified from the prospective neurosurgery patient database. All patients underwent surgery for VNS lead revision diagnosed by X-ray visualization or by high lead impedance greater than 5,300 ohms or less than 600 ohms. Infectious indications for surgery were excluded from review as this may have impacted VNS replacement, though a similar technique could be applied in this cohort. Demographic, clinical, and device data were collected. The primary outcome was perioperative adverse events with complete removal and replacement of the lead from the vagus nerve. Perioperative adverse events included new neurological deficit, surgical site infection, vocal disturbance, dysphagia, or other untoward postoperative outcomes in the perioperative period. Device current at follow-up after VNS revision was available for a subset of these patients.

Operative technique
All patients underwent complete removal of the existing implant leads from the vagus nerve and placement of new leads on the same segment of the vagus nerve in one operation. Patients were positioned supine, with a roll placed under the shoulders, and the head slightly extended and turned right. Sterile approach to the VNS implant at the left neck and chest wall used blunt and sharp dissection to the implant, with no unipolar electrocautery during the entirety of surgery. If there was no radiographic confirmation of a broken lead, the existing VNS lead is disconnected from the generator and reconnected, to check the impedance. If impedance remained over 5300 ohms or less than 600 ohms, the VNS lead was revised. At the neck incision, the strain relief loop of the VNS lead was identified and followed deeper, toward the carotid sheath. At this point, the traditional teaching was to truncate the VNS lead, less than 2 cm, without exposure of the vagus nerve, leaving the coils on the nerve. With the current technique, the previously placed proximal lead wire was further exposed, with two helical electrodes and helical tethering anchor identified on the vagus nerve. Initially, if only one or two coils were first visible, continued sharp dissection was employed to further dissect out the entire series of helical coils. The most important instrument and technique of the surgery were the use of fine pointed scissors to develop the plane immediately below the polymer material of the coil, hugging the undersurface of the coil and staying in its plane, parallel to the nerve spreading the scissor tips gently while withdrawing them to develop that plane, taking care to stay parallel to the nerve while gently dissecting to avoid injury to the jugular vein and carotid artery. The coils then loosen around the vagus nerve, leaving an intact vagus nerve. This process was repeated for all three coils, allowing for the complete removal of the lead from the nerve which preserving the segment of the exposed vagus nerve. In the authors' experience, there is no significant fibrous ingrowth between the polymer of the helical coils and the vagus nerve. Additionally, the plan for replacement is clear as the leads demarcated this after removal of the three helices. After removal, a new lead is tunneled, placed on the same exposed segment of the vagus nerve, or on a more proximally dissected portion of the vagus nerve when possible if the scarring was prohibitive to replacement, and then connected to the generator using conventional technique for implantation in the same procedure [17,26]. Intraoperative images of the applied technique for VNS lead removal used in this study are demonstrated in Fig. 1. Figure 2 depicts an illustration of the dissection technique, using fine, sharp scissors, dissecting parallel to the nerve, in the plane between the coil and the nerve, sharply dissecting and spreading the scissor tips to free the nerve from the coil.

Results
Fourteen patients underwent VNS lead removal and samesite replacement in the study period. Mean age at stimulator placement was 7.6 years (SD 3.5, range 4.5-13.4) ( Table 1). The most common etiologies of epilepsy were genetic (5, 36%) and cryptogenic (4, 29%). Initial VNS placement was performed at a different institution in 3 patients (21%). Four patients (29%) underwent preoperative video electroencephalography (vEEG), and 11 (79%) underwent preoperative magnetic resonance imaging at our institution, the findings of which are detailed in Table 1. Lead revision surgery was performed at a mean of 6.0 years (SD 3.8, range 2.1-11.7) following initial VNS placement. Reasons for revision were VNS lead breakage or malfunction, diagnosed by X-ray visualization or by high lead impedance greater than 5300 ohms or less than 600 ohms. Implantable pulse generator replacement was performed concurrently with lead replacement in 11 patients (79%). All VNS implants were from LivaNova, Inc. (Clear Lake, TX, USA) and included a variety of models from Model 102 to Sentiva generators.
All patients tolerated the procedure well. Intraoperatively, the vagus nerve was anatomically preserved with no discontinuity or concern for injury in all cases. There were no perioperative complications, including hematoma, infection, or new neurological deficit, following surgery in these patients. Voice and swallowing function were subjectively unchanged postoperatively for all patients. VNS settings were available for 5 of the 14 patients (Table 2). Seizure outcome data was available for five patients, ranging from 25 to 100% total seizure reduction compared to baseline preoperatively. Device data was available for five patients ( Table 2). Two patients required slightly increased current (adjusted from 1.5 to 1.75 mA in both patients). Another two patients required a reduced current (adjusted from 2.0 to 1.75 mA and 2.0 to 0.5 mA). Current remained unchanged in the remaining patient. The mean length of follow-up was 7.9 years for all patients (SD 3.5, range 3.1-13.7).

Discussion
We describe an operative technique and related perioperative outcomes for VNS lead removal and same-site reimplantation in children. We found that lead replacement can be performed without complication and suggest feasibility of same-site VNS lead revision. In the several patients with device data, we found that lead replacement on the same-site did not affect device settings in a meaningful way. To our knowledge, this is the first study to report on stimu- lation outcomes following same-site VNS replacement in a pediatric cohort, and while our data is limited, we believe this study demonstrates both feasibility of this technique but highlights an important metric for future study.

Importance of removing the VNS leads
Complete removal of VNS leads is preferable, when possible, compared to truncation of the retained lead and removal of the distal lead and IPG as the presence of retained leads may impact the patient's ability to undergo certain MRIs, limiting additional neuroimaging studies [35]. Moreover, preservation of the vagus nerve via same-site reimplantation of VNS leads is desirable in pediatric patients, who may require further VNS lead revisions over their lifetime. New VNS lead placement on an alternate, unoperated segment of the vagus nerve may prove to be difficult with less nerve length available for surgical dissection and exposure. However, the complete removal of VNS leads is not widely explored in the literature and published operative techniques for lead removal vary greatly.

There is limited evidence regarding the optimal approach for VNS lead revisions
VNS lead removal has been infrequently described in pediatric cohorts compared to adults. Agarwal et al. reported lead revision and replacement in a 20-patient pediatric Fig. 2 Illustration of dissection of the VNS lead from the vagus nerve. This illustration depicts the relevant surgical anatomy including superficial landmarks, such as the medial border of the sternocleidomastoid, and deeper anatomy of the carotid sheath (green), carotid artery (red), jugular vein (blue), and vagus nerve (yellow). The insets depict the fine, sharp scissors which are used to dissect immediately under the helical coil, parallel to the nerve, sharply, then spreading to free the coil from the nerve  [30,33,36,37] and monopolar electrocautery [31,34,39] are the most commonly described surgical techniques for VNS lead removal. As fibrotic scar tissue overlying the lead coils can form, electrocautery has been used to separate the VNS lead from scar tissue. However, concerns regarding the potential for thermal nerve damage due to electrocautery exist [33,39]. Sharp and blunt dissection methods, such as the use of sharp, fine scissors as employed in our approach, avoid this potential risk but may require more laborious dissection around neurovascular structures in order to free the VNS lead. Other, sharp dissection techniques include piecewise truncation of the anchoring loops with scissors followed by electrode removal [33] and direct truncation of the lead coils [36]. Espinosa et al. described a related surgical technique to the one used in this study, in which tenotomy scissors were used to pry open lead coils [30]. This is similar to our use of sharp dissection to access the plane between the nerve and the coil with use of gentle dissection via spreading of the fine, sharp scissors, in this plane to uncoil the lead from the nerve. These techniques for VNS lead removal avoid the use of electrocautery near the vagus nerve and allow removal of the lead in one piece, affording ample dissection to remove the lead while preserving that segment of the nerve without thermal risk to the nerve. Few studies report on outcomes of same-site VNS lead reimplantation, and even fewer explore this in pediatric populations. Dlouhy et al. reported outcomes of VNS lead revision and same-site replacement in a cohort of 24 adult patients using a sharp dissection technique for lead removal, noting that baseline seizure control was attained in all but one patient receiving 3-month follow-up [37]. Ng et al. described VNS lead removal and same-site replacement in 5 pediatric patients using electrocautery with an ultra-sharp monopolar tip to dissect scar tissue followed by coil removal with non-toothed forceps but did not report outcomes [34]. Same-site reimplantation of VNS leads was also performed in both adults and pediatric patients by Aalbers et al., with removal performed by cutting the helical coils with straight or blunt-tipped scissors [36]. One case of postoperative vocal cord paralysis was encountered, and patients reported baseline or improved seizure control on a follow up survey. [36]

Potential implications in stimulation parameters
To our knowledge, this is the first study examining the impact of lead revision with same-site replacement of a new VNS lead on stimulation parameters. We found that stimulation parameters for lead replacement on the same nerve segment demonstrated minimal changes to pre-revision settings, suggesting that similar seizure control from VNS can achieve minimal changes to postoperative stimulation parameters after lead revision using this technique. While our sample size was too small, precluding assessment of statistical significance, this suggests that same-site insertion does not adversely impact device settings, carrying potential implications regarding battery usage over time and frequency of pulse generator replacements in these children. Any theoretical scarring that might be present on the nerve segment does not impact the ability to effectively neuromodulate the vagus nerve at that same-site. However, larger scale, prospective studies are warranted to further evaluate the impact of VNS lead revision and replacement on the same segment of the nerve impacts device settings.

Alternate approaches to VNS lead revisions
As aforementioned, there are a variety of techniques applied to manage VNS lead revisions; these include truncation with replacement on a different segment of the nerve, removal with replacement on a different segment of the nerve, and removal with replacement on the same segment of the nerve. The surgical techniques to accomplish these dissections and revisions are performed in a variety of manners. However, a potential alternative to lead removal and replacement, not explored in our study, is repair. Ralston et al. described in situ repair of lead insulation in a cohort of 4 pediatric patients who presented with low lead impedance, with follow-up time ranging from 4 to 7 years [40]. In situ lead repair involved placement of silicone catheter tubing around the damaged lead and secured with cyanoacrylate glue. Repair of lead insulation may be advantageous as it avoids the risks of reoperation directly on the vagus nerve, minimizing manipulation of the nerve. However, this technique is not applicable in cases of lead fracture, and long-term durability and outcomes of the repair remain to be evaluated. While not explored in our cohort, this surgical repair is yet another tool to consider when approaching these cases.

Limitations
There are several limitations to this study. The cohort is small in size with only 14 patients, limiting our ability to perform analytic statistics. The retrospective design over years also limits the quality of the data able to be obtained, as some records were incomplete or data was inconsistently available, introducing potential bias. Additionally, data regarding device settings was unavailable for all patients due to the retrospective nature of data collection. A prospective study may standardize the outcomes assessed in each subject to ensure more complete assessment of outcome measures. Finally, as this study included patients from a single institution, cared for by two surgeons, the generalizability of our findings may be limited due to practice variation and surgical experience. While we found that VNS lead removal and same-site replacement is feasible in children with refractory epilepsy, further characterization of the optimal technique, postoperative outcomes in terms of associated risk and seizure outcomes, and impact on device settings and usage over time is needed. Larger prospective studies may help to further elucidate the optimal lead removal and replacement technique and the impact of same-site replacement on neuromodulation abilities. A future study exploring stimulation parameters and device settings by replacement technique as well as more large-scale studies may shed further light into the benefits of this technique compared to others. Furthermore, long-term follow-up may afford further insights into the implications of lead removal and revision in the child's lifetime.

Conclusions
We describe our technique for VNS lead removal using sharp scissor blunt dissection without electrocautery for removal and same-site replacement of VNS leads in children. We found that VNS removal and same-site replacement was feasible and safe in this single center cohort, with no postoperative adverse events or untoward impact on stimulation settings. A further study of same-site lead replacement as well as its impact on device usage and settings over time is warranted.