Childhood epilepsies are often characterized by a wide range of seizure types and accompanying comorbidities such as mental impairment or developmental delay (MI/DD), autism spectrum disorders, and language disorders, often making treatment difficult. However, advances in our understanding of the underlying mechanisms that result in seizures and epilepsy syndromes have also led to advances in epilepsy treatments. Traditionally, the two primary treatment modalities used to control seizures have been mono- and polytherapy with antiepileptic drugs (AEDs) and epilepsy surgery. Among pediatric patients, the ketogenic diet is used among a small number of children (see Chapter 70).

The European Union (in 1994) and the U.S. Food and Drug Administration (FDA) (in 1997) approved the first nonpharmacologic treatment for epilepsy—vagus nerve stimulation (VNS) therapy. Since this time, several neurostimulation devices have been approved in Europe and the United States, although the exact indication and devices approved have differed between the two continents (Table 78.1).

In Europe, only trigeminal nerve stimulation (TNS) and VNS therapy are approved for use in children; in the United States, only VNS therapy is approved for use in children. To date, VNS therapy accounts for almost all of the worldwide use of neurostimulation devices in the pediatric age range, and has the most extensive scientific literature supporting its use; thus, VNS therapy will be the focus of this chapter, although the other devices and their indications will be mentioned at the end.

VAGUS NERVE STIMULATION THERAPY

Although VNS therapy is not currently approved in the United States for children younger than age 12 years, studies indicate that success with VNS therapy can be achieved independent of patient age and seizure type or syndrome.

Reports indicate that VNS therapy may have unique benefits for pediatric patients (aged less than 18 years), including improvements in quality of life resulting from the lack of pharmacologic interactions known to impair development and success at reducing seizure frequency and severity among patients with age-related and difficult-to-control syndromes such as Lennox–Gastaut syndrome (LGS). The American Academy of Neurology has recently updated its evidence-based guideline for VNS (1). This lists children, and specifically LGS, as appropriate candidates. This section outlines both the safety and effectiveness of VNS therapy among pediatric patients with epilepsy, as discerned from current treatment practices and reports in the literature.

THE VAGUS NERVE STIMULATION THERAPY SYSTEM

VNS was the first nonpharmacologic therapy approved by the FDA for the treatment of seizures. The treatment, which attenuates seizure frequency, severity, and duration by chronic intermittent stimulation of the vagus nerve, is intended for use as an adjunctive treatment with AED therapies. As of July 2015, more than 100,000 patients with epilepsy have been implanted with the VNS therapy system worldwide, with approximately 30% of those patients being younger than age 18 at the time of their first implant. Approximately 33% to 45% of patients receiving VNS therapy experience at least a 50% reduction in seizure frequency with no adverse cognitive or systemic effects (2,3). Moreover, clinical findings indicate that the effectiveness of VNS therapy continues to improve over time (3–9), independent of changes in AEDs or stimulation parameters (10). Also notable is the fact that tolerance, which is often accompanied by a reduction in efficacy for many treatments, does not appear to be a factor with VNS therapy, even after extended (>10 years) periods of time (3). Response to VNS therapy may be delayed for some patients (9,11). As a result, the long-term safety and effectiveness seen with this treatment have made VNS therapy a mainstream treatment option for a broad range of epilepsy patients, including children and adolescents. VNS therapy is second only to AED therapy as a treatment category for childhood epilepsy in the United States.

TABLE 78.1

The VNS therapy system consists of an implantable pulse generator and bipolar VNS therapy lead, a programming wand with software, a tunneling tool, and a handheld magnet. The original systems consisted of Model 100 and Model 101 (these are no longer distributed). However, in June 2002, Cyberonics, Inc. (Houston, Texas) introduced a new generator model, the Model 102 system, which is thinner (6.9 mm), lighter (25 g), and has less volume (52.2 mm in diameter) than the previous models (12). Subsequently, Cyberonics has introduced several other models (Table 78.2).

The Demipulse (Model 103) is significantly smaller, making for an improved cosmetic appearance and increased comfort in young children; however, the battery life is shortened compared to the other models. The newest version, Model 106, allows automated or on-demand magnetic stimulation (using ictal tachycardia as a surrogate marker for seizure onset) (13–15). This model (Aspire SR) has the potential to shorten seizure duration and further improve quality of life. Model 102 has a single-rather than a dual-pin lead, making it easier and faster to implant than the previous models. The average battery life for the Model 105 generator is approximately 7 to 10 years with normal use (12). Increases in stimulation intensities or frequency will decrease battery life.

TABLE 78.2

The magnet provided to patients as part of the VNS therapy system allows for on-demand stimulation, which has the potential to abort seizures, either consistently or occasionally, among some patients or caregivers who are able to anticipate the onset of their seizures (13,16–19). The additional stimulus train that results when the magnet is held over the generator is typically stronger than the programmed stimulus parameters. This added ability of on-demand stimulation provides a greater sense of control for patients and their caregivers over their disorder, which can help improve how they perceive their quality of life. The magnet also allows temporary interruption of stimulation if needed, particularly when singing or playing wind instruments or during speaking engagements. However, stopping the stimulus should be done sparingly and with care, as doing so creates the potential risk of seizure occurrence.

Implantation Procedure

The implant surgery is most often performed under general anesthesia and typically lasts under 1 hour (5). The pacemaker-like generator device is generally implanted in the subcutaneous tissues of the upper left pectoral region, with a lead then run from the generator device to the left vagus nerve in the neck, where it is attached by a coiled electrode (Figures 78.1 and 78.2) (20–22). Two incisions are made during the procedure: one in the chest to create the generator pocket and the other along a fold in the neck to expose the vagus nerve for placement of the electrode (Figures 78.3 and 78.4). A loop of lead wire is coiled beside the generator to allow for strain relief and patient growth.

FIGURE 78.1 VNS lead wire prior to placement on the left vagus nerve. The cathode electrode is placed proximally (right side of picture), then the anode electrode, and then the anchor tether (caudal; left side of picture).

FIGURE 78.2 Lead wire starting to be placed on the left vagus nerve.

The procedure is well tolerated in both children and adults (23–25) and is usually performed as outpatient surgery; however, in some cases, patients may be kept in the hospital overnight for observation. Minor modifications of the location are required for the Model 106 generator (26). The device is often turned on in the operating room or in the office immediately after surgery, generally with a low initial setting of 0.25 mA (Figure 78.5) (27). The programming wand (Figure 78.6) is used at follow-up visits to check and fine-tune the stimulation settings according to patient comfort and level of seizure control. Instructions concerning care of the incision sites, magnet use, and necessary follow-up visits are given to patients and their families before the patient leaves the hospital.

FIGURE 78.3 Implantation of Model 102.

FIGURE 78.4 Neck incision (right) after final closure with Durabond; chest incision after final closure (bottom left).

The length of battery life for the VNS generator is dependent on the device model implanted and the stimulation parameters used (28). Often an increase in seizure frequency or intensity suggests the clinical end of service (28). Other indications of battery failure include a sudden stop in sensing stimulation or unexpected changes in stimulation. Once a generator reaches its end of service, another surgery is required to replace the generator. The entire generator is replaced rather than just the battery so as to prevent opening the hermetically sealed titanium case of the generator, which could lead to a rejection reaction (29). The generator-replacement surgery typically lasts approximately 10 to 15 minutes and is performed on an outpatient basis. Because the leads remain untouched during a generator replacement, only one incision is needed. There are now two lead-wire options available, one a single pin and one a dual pin. It is critically important to be aware of which lead wire the patient has in place so the correct generator is available in the operating room. Generator replacement is recommended and preferred by patients before the battery is completely depleted so as to prevent an interruption in treatment (28). Generator replacement performed prior to the end of service allows the same settings to be programmed into the new device that were being used for the old device, preventing interruption of service (eg, the old device is interrogated in the operating room and the parameters are recorded; then the new device is connected and the same parameters programmed into the new device). A 12-year follow-up study showed that multiple device replacement surgeries are well tolerated (3). If replacement is performed after the battery was allowed to reach its end of service, then programming must be reinitiated at 0.25 mA and titrated up, similar to a new generator placement. For patients who are not felt to have obtained benefit from VNS, it has been suggested that stimulation parameters be tapered down over 6 months (by decreasing the output current by half, following up in 3 to 4 months, and, if no worsening in seizure control, setting the output to 0 mA and observing seizure control again over at least 3 to 4 months) before reaching the end of service to allow for decisions about continuing VNS therapy (or performing explantation of the device in the future) to be made. This allows the clinician and the family to know exactly when the device was turned off, and if there is any clinical worsening to reinitiate therapy while scheduling a prompt generator replacement. If there is no clinical change, the family and physician can have the device and complete lead wire removed in the future if it continues to not be used (ie, there is no urgency to do this, and it is better to make sure it will not be used again before rushing to removal).

FIGURE 78.5 Intraoperative use of the handheld computer and programming wand to initiate stimulation.

FIGURE 78.6 A programming wand is held by the patient over the device while a physician checks and/or adjusts stimulation parameters using a handheld computer.

Potential Complications

Although the implant surgery is a relatively simple procedure that is safe and well tolerated by the vast majority of VNS therapy patients, complications can arise (30). One possible risk resulting from the implant surgery is an infection at the implant site (31–34). This risk may be increased in the pediatric population because young children or patients with neurocognitive disorders may tamper with the wound before the incision has had time to heal properly (1,35–37). Such infections can be treated with antibiotics but typically lead to explantation of the device if antibiotic treatment is not effective or if tampering continues (38,39). Stimulator pocket infections have been relatively uncommon among the pediatric population (38,40). No infections were observed in a recent study of 36 children aged younger than 18 years who were followed up for an average of 30 months (40).

The routine lead test performed during surgery also has resulted in reports of bradycardia and asystole in a small number of patients (~0.1%) (41–44). Neither of these cardiac events, however, has occurred after surgery during day-to-day treatment with VNS therapy or in children; these events are usually transient and self-limiting and are rarely of clinical significance (41–44). Vocal cord paresis, although rare, can be caused by manipulation of the vagus nerve during the implant procedure, but such paresis is most often transient and less common in children compared to adults (45). A lower neck incision and a submuscular pocket for the generator may also minimize laryngeal complications (31,46,47). On the whole, the surgery required with VNS therapy is much less invasive and generally better tolerated than other traditional epilepsy surgeries. Although side effects associated with the surgery cannot be avoided completely, they can be minimized with the correct technical procedure (40). In addition, the implant surgery is not associated with any performance or cognitive impairments and can be reversed if the treatment is not effective.

Alternative Generator Placements

Depending on the circumstances of the patient, alternative generator placements have been reported with successful results. Le et al (35) successfully used an interscapular placement of the generator to reduce the risk of wound tampering among pediatric patients with cognitive delay who may be prone to tampering with the wound. Of the nine patients with an interscapular generator placement, none required explant; no discomfort or changes in daily routine (eg, sleeping positions) were reported; and the effects of VNS therapy on seizure reduction and quality of life were consistent with the results seen with the traditional generator placement. An infraclavicular placement also has been used for developmentally delayed children, as well as for young women and children with small muscular mass (48,49). Three children have had a device reimplanted on the right side rather than the left after undergoing explantation of a VNS device owing to infection (50). All three patients had been deriving benefit from the left-sided implant, but reimplantation of another device on the left side was not pursued because of the potential of inflicting nerve damage. The right-sided implantations were effective in once again reducing seizure frequencies, but a difference was noted in the level of seizure control achieved between the right- and left-sided implants. Although right-sided implantations are not recommended by the device manufacturer because of possible cardiac side effects, no cardiac events presented among these three children; however, two of the three children did experience transient respiratory events.

Stimulation Parameters

VNS therapy “dosing” is defined by five interrelated stimulation parameters (Figure 78.7): output current (measured in milliamperes), signal frequency (Hz), pulse width (µs), signal on time (s), and signal off time (s/min). The output current, signal frequency, and pulse width define how much energy is delivered to the patient, with the combination of settings for these three parameters being analogous to the size or dose of a pill (51). The signal on and off times constitute the duty cycle (ie, how often the energy is delivered) and are analogous to the dosing schedule for drug therapy. An optimal dose–response relationship for VNS therapy, however, is elusive, in part because of the interindividual variability between patients and in part because of the number of parameters involved in regulating the dose.

Standard parameter settings, as determined from the clinical trials and outlined by Heck et al (52), range from 20 to 30 Hz at a pulse width of 250 to 500 ms and an output current of 1.50 to 2.5 mA for 30 s on time and 5 min off time (Table 78.3). This is consistent with mathematical modeling simulation studies of the vagus nerve, which demonstrate optimal activation of the myelinated fibers at these settings (51). Initial stimulation is set at the low end of these ranges and slowly adjusted over time and within the safety limits (Figure 78.8) on the basis of patient tolerance and response. Patients should be closely monitored during the dose adjustment phase of VNS therapy, typically every 2 to 4 weeks for the first 2 to 8 weeks following generator implantation. Once a patient responds to a tolerated dose, further parameter adjustments are performed only as clinically necessary. However, routine assessment of lead-wire integrity and generator function should be performed.

FIGURE 78.7 Stimulation parameters (all duty cycles except low output [≤10 Hz]).

Response to VNS therapy has been shown to be age-dependent; therefore, VNS stimulus parameters may need to be adjusted differently for the pediatric patient (53). Several studies indicate that pediatric patients may require higher output currents (Table 78.1) than those used in adult patients to reach a therapeutic dose (2.0–2.5 mA compared to 1.0–1.75 mA, respectively), particularly when lower (≤ 250 µs) pulse durations are used (48,52,54). Additional reports indicate clinically significant responses with low-stimulation intensities (1.25–1.50 mA) (3,6,23). A multicenter, randomized trial looking at device parameter efficacy showed that various duty cycles were equally effective (55). Some reports among children with severe epileptic syndromes showed an increase in seizure frequency and severity at higher output currents (11,56). Functional magnetic resonance imaging (fMRI) in humans also suggests that pulse width is an important variable in producing brain effects (57). Optimal parameter settings for specific ages and seizure types or syndromes, if they exist, have yet to be defined.

TABLE 78.3

FIGURE 78.8 Safety ranges for VNS therapy stimulation parameters.

Mechanisms of Action

The mechanisms of action of VNS therapy are not fully understood, but they are believed to be manifold, owing to the diffuse distribution of vagal afferents throughout the central nervous system, and are distinct from those of traditional AED therapy (58,59). Studies suggest that altered vagal afferent (not efferent) activities resulting from VNS are responsible for mediating seizures (59,60). Such altered activities have been recorded in both cerebral hemispheres. Rat studies indicate that VNS activation of the locus coeruleus may be a significant factor for the attenuation of seizures (61–64). Human imaging studies also implicate the thalamus in having an important role in regulating seizure activity (65–67). The exact antiseizure role of the thalamus is likely to be complex, however, because of the diffuse connections of the thalamus throughout the brain (65). Additionally, the reticular activating system, central autonomic network, limbic system, and noradrenergic projection system all may play a role in the antiseizure mechanisms of VNS therapy (59,68–72).

PET imaging studies in humans show that the VNS-induced changes in cerebral blood flow (CBF) and synaptic activity vary over time (71). Widespread alterations in CBF activation during acute VNS were much more restricted after prolonged VNS, indicating that those sites of persisting VNS-induced changes may reflect the antiseizure actions of VNS therapy (71). During chronic VNS, no new sites of CBF alterations that were not also affected acutely were observed (71). Additional human imaging studies using various methods, including fMRI and single photon emission computed tomographic (SPECT) techniques, have produced similar findings in both acute and chronic studies (71). These imaging findings, coupled with the clinical findings that the effectiveness of VNS therapy continues to improve over time, seem to indicate that rapidly occurring subcortical effects rather than rapidly occurring cortical effects may be more important in the VNS antiseizure mechanism (71). It is believed that rapidly altered intrathalamic synaptic activities as well as other mechanisms likely occurring independently of thalamic activation constitute the therapeutic mechanisms of VNS (71). Electroencephalogram (EEG) observations also suggest that unilateral rather than bilateral abnormalities may show more benefit from VNS (3). Other human studies show that some antiepileptic mechanisms affected by VNS are either modulated by or are the reflection of EEG changes, although the effect of VNS on the EEG background remains uncertain (73). Gamma-aminobutyric acid (GABA) receptor (GABAA) plasticity may contribute to the ability of VNS to modulate the cortical excitability of brain areas associated with epileptogenesis (74). Finally, changes of tryptophan metabolites assessed in cerebrospinal fluid of children responding to VNS suggest treatment may have an anti-inflammatory property (75). Like medications, many different actions probably contribute to the efficacy of VNS therapy; therefore, determining whether there is a single mechanism that is most important may not be possible.

Animal Trials

VNS therapy was developed on the basis of early findings of neuroinhibition in the regulation of emesis, as well as changes in EEG activity resulting from vagal stimulation (76–78). Studies among animals showed that VNS therapy worked both acutely to abort seizures and chronically to control seizures (79,80). In addition, VNS was effective beyond the period of active stimulation and across a wide range of seizure types and severities, thereby indicating the potential for long-term, phase-dependent seizure control. Tests of VNS therapy in the traditional animal models used to test AED efficacy, including rat, canine, and monkey models (80–89), further indicated that VNS therapy, like AEDs, may be effective for multiple seizure types. The clinical trials that followed proved the safety and efficacy of VNS therapy for controlling seizures, with few and mild side effects that were generally related to stimulation intensities.

A more recent PET imaging study in rats revealed differences between acute and chronic changes in glucose metabolism during VNS, which may reflect cerebral adaptation to VNS (90).

These findings are in line with clinical findings and other animal and human studies that show improvements in seizure control and changes in CBF over time (4–6,90,91). Other animal studies also have shown that VNS may facilitate the recovery of function following brain damage, both in the rate of recovery and in the final level of performance reached postinjury, as well as enhance memory storage processes (92–95). This enhancement of neural plasticity is believed to occur by the ability of VNS to enhance norepinephrine release throughout the neuraxis, but further studies are needed to explore the mechanisms behind these effects more fully (93,96). Improvements in behavior among rats with induced brain damage also are suggestive of a neuroprotective effect of VNS, which may also help to reduce the behavioral deficits associated with seizures among humans receiving VNS therapy as part of their antiepileptic treatment (92).

SEIZURE EFFICACY

Clinical Trials

A series of acute-phase studies with long-term follow-up data proved the safety and efficacy of VNS therapy for the treatment of refractory epilepsy and thereby led to its approval by the FDA in 1997. Results from two randomized, placebo-controlled, double-blind trials (E03 and E05) were pivotal in demonstrating the antiseizure effect of VNS therapy. Patients in the 14-week acute E03 study (n = 113) who received therapeutic doses of stimulation (high; 30 Hz, 500 µs, up to 3.5 mA for 30 s every 5 min) showed a significantly higher median decrease in seizure frequency of 24.5% compared to only a 6.1% median decrease among patients receiving nontherapeutic levels of stimulation (low; 1 Hz, 130 µs, ≤ 3.5 mA for 30 s every 90 min; P = .01) (97).

Seizure frequency reductions of at least 50% were reported for 31% of patients in the high-stimulation group and for 13% in the low-stimulation group (P = .02). In the acute E05 study, the median reduction in seizure frequency was 27.9% for the 94 patients in the high-stimulation group and 15.2% for the 102 patients in the low-stimulation group at 3 months (P = .04) (98).

Long-term follow-up of patients in the acute E03 and E05 studies showed that the effectiveness of VNS therapy was cumulative. After the acute phase of these studies, patients receiving low stimulation had their stimulation levels titrated to therapeutic (high) settings, and all patients were then followed for an additional 12 months of treatment. For the 100 patients in the E03 open-label extension study treated for the additional 12 months, the median reduction in seizure frequency increased to 32% at 12 months from only 20% at 3 months (6). For the 196 patients with evaluable seizure data in the E05 trial, the median reduction in seizure frequency was 45% after an additional 12 months of treatment in the prospective, long-term extension study (5). At this same time point, 35% of patients had at least a 50% reduction in seizure frequency, and 20% had at least a 75% reduction. These seizure frequency reductions were sustained over time.

Pediatric Outcomes

Although the controlled clinical trials did not focus specifically on the pediatric patient, the children and adolescents included in one of the five clinical studies (E04) responded at least as favorably as the adults (18,99). Of the 60 pediatric patients included in the E04 open, prospective study, 16 were younger than age 12 (mean age, 13.5 years). At 3 months, the median reduction in seizure frequency was 23% (n = 60); for the 46 patients with follow-up data available at 18 months, the median reduction was 42%. The results, although in a much smaller group, were similar for the patients aged 11 years and younger, indicating that age does not seem to be a factor in the effectiveness of VNS therapy to control seizures. Moreover, data from other pediatric study experiences with VNS therapy indicate that younger patients may have a higher tolerance and more effective response than adult patients (18,23,99–102).

The first large study to evaluate the effectiveness, tolerability, and safety of VNS therapy among pediatric patients was a six-center, retrospective study of 125 patients aged 18 years or younger (41 patients aged less than 12 years) (27). This study showed greater reductions in seizure frequency than those found in the pediatric subgroup of the E04 clinical trial, with a median reduction in seizure frequency at 3 months of 51.5% (range, −100% to +312%; n = 95) and 51.0% at 6 months (range, −99.9% to +100.0%; n = 56). A similar response was reported at 6 months for children aged less than 12 years (median seizure frequency reduction of 51%; n = 20). These reductions did not differ between patients with different seizure types. Subsequently, three large series (N = 100, N = 141, and N = 347) have reported on the efficacy of VNS in children. These all show response rates equal to or better than those seen in adult patients (30,103,104).

The largest study to date to evaluate the effectiveness, quality of life, and safety of VNS therapy among pediatric patients was a European 11-center, retrospective study of 347 children (aged 6 months to <18 years) (104). At 12 and 24 months after implantation, 37.6% and 43.8%, respectively, of patients were responders (had a ≥ 50% seizure reduction).

Special Patient Populations

Although few prospective or controlled trials have been performed among pediatric epilepsy patients, the number of young patients receiving VNS therapy across the United States and Europe is growing. Observations of pediatric patients with age-related or specialized syndromes receiving VNS therapy indicate that this treatment is safe and effective across a broad range of seizure types and syndromes, independent of age. Table 78.4 (3–6,8–10,18,21,24,27,30,37,54,97–101,104–127) shows the epilepsy syndromes, seizure types, and associated conditions in which VNS therapy may be helpful. Additionally, VNS therapy also seems to be a palliative treatment option for patients who have failed cranial surgery (Figure 78.9) (112,116,128).

Lennox–Gastaut Syndrome, Infantile Spasms, and Ring Chromosome 20 Syndrome

LGS and infantile spasms are rare and difficult-to-treat epileptic disorders. These conditions are also accompanied by neurologic comorbidities that can be exacerbated by the cognitive adverse side effects typically associated with drug therapy. Although limited data are available for children receiving VNS therapy for the treatment of infantile spasms (129), recent retrospective studies of the efficacy of VNS therapy among patients with LGS have shown some success in reducing seizure frequency without adverse side effects (11,49,54,100,104,106,118,119,121,122,124,126,130,131).

The largest retrospective study of LGS patients receiving VNS therapy was done by Frost et al (100) on 50 patients from six centers (median age at implant was 13 years; range, 5–27 years). This study showed that median reductions in seizure frequency at 1, 3, and 6 months of VNS therapy were 42%, 58.2%, and 57.9%, respectively (n = 46 who had complete seizure data available). Seizure reductions at 6 months by type showed an 88% decrease in drop attacks and an 81% decrease in atypical absence seizures. At 3 months, a 23% decrease was seen in complex partial seizures. In addition, improvements in quality of life with minimal and tolerable side effects from both the surgery and therapy were reported for this patient population. The most notable change in quality of life was an increase in alertness reported for more than half of the patients. Previous corpus callosotomy was not a contraindication for VNS therapy among this patient population, with the five patients who had undergone such surgery showing a 69% reduction in seizure frequency at 6 months. However, the one patient with a previous lobectomy surgery did not show a change in seizure frequency with VNS therapy. Age also was not a contraindication, because those patients aged less than 12 years showed similar response rates to the group as a whole.

TABLE 78.4

JME, juvenile myoclonic epilepsy.

FIGURE 78.9 Suggested treatment sequence flowchart for patients with epilepsy.

A smaller, longer term study on the behavior of 19 patients with LGS receiving VNS therapy showed no deterioration from baseline in either quality of life or cognitive measures after 24 months of treatment (137). In addition, a positive increase in cognitive measures indicated a gain in mental age of 4.2 months at follow-up. These findings were independent of seizure response to VNS therapy, indicating a potential additional treatment benefit from VNS therapy to patients with LGS. Two-year follow-up of 19 children (aged 7–18 years) with LGS or Lennox-like types of epilepsy, mental retardation, and multiple seizure types, along with a high seizure frequency, showed that four patients had a seizure frequency reduction of at least 50% (49). One patient remained seizure-free 2 years postimplantation. Atypical absence seizures responded better to VNS than other seizure types; moreover, higher baseline seizure frequency, the lowest quantity of interictal epileptic activity, and the highest baseline mental level seemed to predict a higher response rate. A mild improvement in mental age, though modest, and a positive effect on behavior that was independent of seizure control were seen over time. However, unlike some studies that show increased effectiveness over time with VNS therapy, the duration of treatment in this small study did not appear to increase VNS efficacy among this group of patients.

A long-term study followed 30 patients with LGS for a mean observation time of 52 months. The median reduction in seizure number was 60.6%; the best response was observed for atonic seizures (80.8% mean reduction) and the least effect was seen in generalized tonic–clonic seizures (57.4% mean reduction) (124). Additionally, improvements in alertness, postictal phase, and reductions in AED use were seen.

A case study by Buoni et al (11), however, indicated that the ability of VNS to reduce seizure frequency among patients with LGS may require extended periods of treatment before positive outcomes are observed. A 22-year-old man with LGS reported no reductions in seizure frequency for 1 year of VNS therapy, although improvements in alertness and appetite were achieved. But by 3 years, the patient’s seizures abruptly disappeared, and his EEG was borderline normal without any changes in drug therapies or stimulation parameters.

VNS and corpus callosotomy are options for treating refractory seizures associated with LGS. No study has directly compared the response rate for these two interventions in an LGS population. Lancman et al performed a meta-analysis of VNS versus corpus callosotomy in the treatment of seizures associated with LGS to answer the question regarding their comparative efficacy (138). Nine corpus callosotomy studies and 17 VNS studies met their inclusion criteria. Corpus callosotomy had a significantly better outcome than VNS for responder rates (> 50% seizure reduction) for atonic seizures (80.0% vs 54.1%) (< 0.05) (138). All other seizure types, as well as the total number of seizures, showed no statistically significant difference between treatment with VNS or corpus callosotomy.

Finally, the American Academy of Neurology updated their evidence-based guideline regarding VNS in 2013 and concluded that VNS should be considered in patients with LGS (1).

Three separate cases of the use of VNS therapy among patients with ring chromosome 20 have been reported in the literature (56,139,140). In 2002, the first report of a girl aged 6 years indicated that VNS may be successfully used to treat the medically refractory generalized clonic epilepsy characteristic of this disorder, which also is marked by developmental delay (140). After receiving VNS therapy, the child became free of seizures and remained seizure-free for 9 months postimplantation. Moreover, the child was reported to have an increased level of alertness and less lethargy, and, after 4 months of VNS treatment, spontaneous vocalization was reported. The achievement of new developmental milestones after the initiation of VNS therapy was encouraging. A second report of a male implanted with a VNS device at the age of 8 years also showed a good response to VNS therapy (56). Seizures began at the age of 5 years, and the patient was experiencing numerous absence and nocturnal tonic–clonic seizures as well as nonconvulsive status epilepticus at the time of VNS implantation. Following initiation of VNS, seizure frequencies were eventually reduced to only occasional nocturnal episodes and some previously lost skills were reacquired, including ambulation, eye contact, social smiling, and improved mood. The third report was a 14-year-old male with ring chromosome 20 syndrome who was also diagnosed with LGS and experiencing a range of seizure types and severe impairment of cognitive functions. However, he did not show similar results to those reported previously (140). Following implantation of the VNS device at the age of 11 years, the child experienced a 50% reduction in seizure frequency. However, tonic seizures during sleep and secondary generalized seizures continued at a rate of 1 to 3 seizures per day. At age 13 years, a corpus callosotomy was performed with no additional benefit in terms of seizure frequency reduction, but some reduction in seizure severity was achieved. An increase in behavioral problems, fear attacks, and visual hallucinations began after callosotomy. These case reports suggest, therefore, that earlier use of VNS therapy among patients with ring chromosome 20 syndrome may be more beneficial.

Mental Impairment/Developmental Delay

MI/DD often co-occur among patients with epilepsy, and the causal relationship between these disorders is complex (141). Both the seizures caused by the epilepsy and the AEDs used to treat the epilepsy are, however, known to potentially exacerbate delays in development, which can complicate treatment among this patient population. The likelihood of additional behavioral and psychiatric disorders, which are estimated to be about seven-fold higher among this population, also further complicates the treatment regimen (141). Moreover, the increased use of polytherapy among this population is indicative of the large number of patients with MI/DD experiencing refractory seizures (142,143). Therefore, VNS therapy may be an attractive treatment option among patients with developmental and behavioral comorbidities in addition to epilepsy, because VNS therapy may reduce the frequency of seizures without the pharmacologic side effects or interactions of additional drug therapy. Another potential benefit is the fact that VNS therapy is delivered automatically, meaning that compliance and caregiver reliance for treatment are minimized, which is particularly attractive for this patient population because many are unable to care for themselves.

Studies of the effects of VNS therapy among patients with MI/DD show success with VNS therapy (37,143,144). A retrospective study by Andriola and Vitale (37) of 21 mildly to severely affected MI/DD patients (age range, 3–56 years; 5 patients <16 years) with a range of seizure types and etiologies showed VNS therapy to be both effective and well tolerated. Seventy-one percent (15 of 21) saw some degree of change for the better in seizure frequency or severity. Of the 16 patients who had known pre- and postoperative seizure data available, all of them had some degree of seizure reduction reported, with 11 (68%) having at least a 50% reduction. One patient with secondary generalized seizures became seizure-free and remained that way at 3 years postimplant. Improvements also were reported by caregivers for many areas of the patients’ functional status, including alertness, mood, and daily task participation. In addition, such improvements were not always associated with decreases in either seizures or AEDs.

Similar findings were found by Gates et al (144) in a retrospective study comparing outcomes of patients receiving VNS therapy living in residential treatment facilities (RTFs) with those not living in RTFs. Despite numerous statistical differences in the demographics and medical histories found between the 86 RTF (age range, 7–59 years) and 690 non-RTF (age range, 2–79) patients, the 12-month responder rates (≥ 50% reduction in seizure frequency) for the two groups were similar at 55% and 56%, respectively. Patients in both groups were reported to have some degree of improvement in alertness, verbal communication, memory, achievement at school or work, mood, postictal period, and seizure clustering, with more improvements reported at 12 months than at 3 months, thereby indicating a cumulative effect of VNS therapy.

A prospective study among 40 patients institutionalized with MI/DD and followed for 2 years of VNS therapy confirmed that VNS was an effective treatment option for this population (143). Most patients (34 of 40) had some reduction in seizure frequency. More notable, however, is the fact that this group experienced fewer epilepsy-related hospitalizations after receiving VNS therapy, and postictal recovery periods were reduced among 75% of the patients. Furthermore, the quality of life for these patients was improved by significant improvements at both 1 and 2 years in attention span, word usage, clarity of speech, standing balance, ability to wash dishes, and ability to perform household chores. Other areas of improvement included an increase in their ability to dress themselves, interact with their peers, express themselves nonverbally, and perceive auditory and visual stimuli.

All of these studies to date indicate that VNS therapy was well tolerated and did not introduce the central nervous system or cognitive side effects that commonly occur when a new AED is added to a treatment regimen (37,143,144). Reported side effects were minimal and manageable with changes in stimulus parameters. In addition, the surgery was less invasive and therefore more tolerable than other epilepsy surgeries. However, wound tampering can be a potential problem with this patient population. One patient in the Andriola et al (37) study was explanted as a consequence of self-inflicted wound infection. This patient was reimplanted and extra measures (extensive bandages over the implant site and additional barriers) were taken to prevent wound tampering until after the implant incisions had healed. As discussed earlier, a second option to prevent wound tampering would be an interscapular placement of the generator (35).

Tuberous Sclerosis Complex, Autism, and Landau–Kleffner Syndrome

A retrospective, multicenter, open-label study of 10 patients (mean age of 13 years) with tuberous sclerosis complex (TSC) receiving at least 6 months of VNS therapy (with a mean of 22 months) found a high response rate to VNS therapy, with 9 out of 10 patients experiencing at least a 50% reduction in seizure frequency (110). More notably, 5 of the 10 patients experienced a more than 90% reduction in seizure frequency. In addition, three patients were reported to be more alert by their parents and teachers, two were reported to have briefer seizures, and one was reported to have less-intense seizures in addition to the reduction in seizure frequency. A patient diagnosed with an autism spectrum disorder in addition to TSC also was reported to have an 80% reduction in injurious behavior after the start of VNS therapy. These results were not countered by any major complications or side effects. In addition, the high response rate of patients with TSC receiving VNS therapy, compared to patients of similar age and seizure frequencies who were also receiving VNS therapy but did not have TSC, indicated that this patient population may be more responsive to VNS therapy. Since this publication by Parain et al, numerous other centers have documented the positive response to VNS therapy seen in patients with TSC (32,119,120,122,123,125,130). A meta-analysis of efficacy and predictors of response to VNS for epilepsy found that tuberous sclerosis was a positive predictor for a favorable outcome (ie, a better response rate than those with unknown etiology) (145).

Preliminary data also suggest that VNS therapy may be effective among patients with epilepsy and either autism or Landau–Kleffner syndrome (LKS), which are both childhood disorders known to co-occur with epilepsy (111). Among six patients with LKS, three experienced a reduction in seizure frequency of at least 50% at 6 months of VNS therapy. Of 59 patients with autism, 58% experienced at least a 50% reduction in seizure frequency at 12 months. More notable, however, were the reported improvements in quality of life, particularly in the area of alertness; 4 of 6 children with LKS and 76% of the children with autism were reported more alert at 6 and 12 months, respectively. Therefore, the benefit of VNS therapy for patients with such disorders may extend beyond or be independent from seizure frequency reductions.

Hypothalamic Hamartomas

A small study of six pediatric patients (≤16 years) with hypothalamic hamartomas and refractory epilepsy indicates that VNS therapy may have the ability to independently improve behavior and, to a lesser extent, decrease seizure frequency or severity in this patient population (146). Three of the six patients experienced some degree of seizure control. However, the immediate and notable improvements in behavior among four of the patients characterized as having severe behavioral problems are of particular interest. Such behavioral improvements were seen independent of seizure control and were dependent on ongoing stimulation. One patient who had the generator turned off for a 2-week period for stereotactic surgery had the injurious and antisocial behavior return in the absence of VNS therapy. Those behaviors once again subsided when stimulation was restarted.

Status Epilepticus

Eighteen pediatric patients (112,113,135) and ten adult patients (135) were identified who underwent treatment with VNS for refractory status epilepticus. In those with generalized seizures, 76% displayed a cessation of status epilepticus after VNS therapy (135). As reported in a case report by Winston et al (112), a 13-year-old boy was implanted with the vagus nerve stimulator 15 days after being admitted to the hospital for pharmacoresistant generalized convulsive status epilepticus. While he was hospitalized, his condition continued to deteriorate despite numerous pharmacologic treatments. The patient also had previously undergone a 90% anterior corpus callosotomy, which had been followed by the return of seizures up to 80% of the preoperative frequency. Immediately following the initiation of stimulation in the operating room, the child’s refractory status epilepticus ceased. Over the next year and a half, the status epilepticus never reappeared, the rate and severity of seizures significantly decreased with little or no postictal phase, and the patient’s neurologic condition, nutritional state, and quality of life all improved.

Another case series presented by Malik et al (113) reports on three children (aged 14 months to 10 years) who also presented with pharmacoresistant status epilepticus and who were emergently implanted with the VNS device. All three children experienced complete resolution of the status epilepticus and continued to have a marked reduction in seizure frequency at their 8-week follow-up visit. The seizure types varied for each of these three patients, with one experiencing atonic, hypomotor, and partial seizures; another atonic, general tonic–clonic, and myoclonic seizures; and the third, multifocal-onset seizures. A 30-year-old man who presented with pharmacoresistant status epilepticus and was placed in a pentobarbital coma also experienced a cessation of seizures and remained seizure-free 19 days postimplantation. These preliminary case reports suggest that VNS therapy should be considered as a nonpharmacologic treatment option among children with pharmacoresistant status epilepticus independently of seizure type.

SAFETY

Stay updated, free articles. Join our Telegram channel

Join