METHODS
A PubMed search was conducted for all English articles up to April 2017 using the
terms “vagus OR vagal OR VNS AND epilepsy, which resulted in 1,394 papers. The query
was then refined by inclusion of all review articles and prospective/retrospective
clinical studies that evaluated VNS efficacy by seizure frequency for at least three
months after implantation.
Historical aspects
Therapeutic options for the management of refractory epilepsy are still limited and
AEDs, resective brain surgery and palliative procedures are possible choices among
all alternatives. Failure of AEDs in controlling seizures after trials with two different
medications in effective doses significantly reduces the chance of improving an outcome
with another drug. Furthermore, despite the development of newer AEDs, there has not
been an increase in efficacy or tolerability of conservative management. In these
patients, brain surgery is still the most promising alternative and should be recommended
when an identified epileptogenic focus can be resected without compromising eloquent
areas. When the investigation is inconclusive, palliative procedures such as disconnecting
surgeries or neurostimulation should be considered. One of the pillars of neurostimulation
is VNS and, in approximately half of implanted patients, this can provide about 50%
seizure reduction[4].
Since 1880, electrical VNS has been used to abort seizures or, at least, to decrease
their frequency and duration. The neurologist James Corning was one of the predecessors
of this procedure and his technique consisted of stimulating the vagus nerve transcutaneously,
in conjunction with carotid compression. This method was initially proposed by Parry
in 1792, with the intent of reducing cardiac output and, consequently, cerebral blood
flow[5]. Apart from side effects, Corning was motivated by the outcomes, but his successors
did not share his enthusiasm and the technique was subsequently abandoned. Nevertheless,
in the 1950s, the interest in vagus nerve stimulation was resumed with animal studies,
as well as its influence on electroencephalography (EEG). With promising outcomes
in animals, a device for human use was designed and first implanted in the 1980s.
In this preliminary study of four patients, two were seizure free, one had 40% improvement
and one did not respond after implantation[6]. Additionally, encouraging results were also achieved in another series of five
patients, published in the same year[7].
These were followed by double-blind, randomized studies, which were extremely relevant
for the establishment of vagus nerve stimulation as an option for refractory epilepsy
treatment and were also considered evidence for therapy approval by regulatory agencies
in Europe and in the United States. Vagus nerve stimulation therapy was approved by
the European community in 1994 and by the American community in 1997 and, nowadays,
more than 70,000 patients have had implants for the treatment of epilepsy or depression[8].
The first controlled, multicenter study (EO3) examined 67 patients above 12 years
of age with partial refractory epilepsy, who underwent VNS implantation and were randomized
to high (20 to 50 Hz) versus low (1 to 2 Hz) frequency stimulation. After 14 weeks,
the mean reduction in seizure frequency was 30.9% in the high frequency group against
11.3% in the low frequency group (p = 0.029), only the former being statistically
significant when compared to the pre-operative status. Moreover, the responder rate
(> 50% reduction in seizures) in the high frequency group was 38%[9]. After a year, a study of 114 patients older than 12 years, including the 67 cited
previously, and following the same protocol, was published describing similar results[10]. Out of these 114 patients, 100 patients were followed in a non-blind study of high
frequency VNS, with 20% reduction in seizure frequency in the first three months,
32% in the last 10 through 12 months and a response rate of 28% and 31%, respectively[11]. In another double-blind multicenter study (EO5), 196 VNS patients older than 12
years with partial refractory epilepsy were randomly assigned to high or low frequency
stimulation and, after three months of follow-up, there was seizure reduction of 27.9%
versus 15.2% in the high and low groups, respectively (p = 0.004). Better scores in
global well-being evaluations were demonstrated (p < 0.001), but no statistical difference
was achieved when comparing the responder rate[12]. After the blind phase, patients were invited to continue follow-up to analyze long-term
outcomes (XE5 study). One hundred and ninety-five patients underwent high frequency
stimulation (including the low frequency group) according to the same protocol of
previous studies. After three months, there was 34% reduction in seizures and, after
12 months, 45% (p < 0,0001). Moreover, when shifted to high stimulation, the low stimulation
group showed better seizure control, confirming the cumulative role of high frequency
stimulation. Nevertheless, it is impossible to conclude that placebo effect (due to
no control group) and variable parameters of stimulation, such as output current and
pulse width, did not impact the results[13].
More recently, closed-loop stimulation, which has already been used for cerebral stimulation
in epilepsy management, has been applied to vagus nerve stimulation. Ictal tachycardia
has been observed in approximately 82% of patients with epilepsy, associated not only
with generalized, but also focal, seizures. Although there is inter- and intra-individual
variability, an increase of 30 bpm or 50% of basal cardiac rate is generally expected
during a seizure[14]. Due to the adversities of continuous noninvasive monitoring, other methods of seizure
detection were investigated and, in February of 2014, a new generator capable of detecting
increases in cardiac frequency related to seizure initiation and trigger stimulation
(responsive stimulation to ictal tachycardia) was approved in Europe. In one implanted
patient, the system was very sensitive but not specific (92% and 13.5%, respectively).
However, after three months of combined stimulation (cyclic and responsive), there
was a reduction of seizure frequency and duration[15]. In 2015, the US - E 37 trial, a prospective and unblinded research for investigating
VNS activated by ictal tachycardia, was also published. Short-term evaluation of 20
patients in the epilepsy monitoring unit after implantation showed that almost 35%
of 89 seizures were treated by the responsive stimulation and 61% of them terminated.
In the long term, the responder rate after 12 months was 50% and the adverse effects
were similar to the previous VNS devices[16]. In another published prospective and multicenter study, short-term evaluation demonstrated
that 40% of seizures were treated by closed-loop stimulation and 58% of them ended.
However, the responder rate after 12 months was 30%, which could be explained by the
parameters used during stimulation, with lower output currents than usual (1.250 mA)[17].
Anatomy
The vagus nerve, also known as ‘X’ cranial nerve, is relatively long and features
sensory and motor innervation, as 80% of its fibers are afferent and 20% efferent[3]. It emerges from the posterolateral sulcus of the medulla in conjunction with the
glossopharyngeal (IX) and accessory (XI) nerves, between the olive and cuneate/gracile
fasciculi. Its efferent fibers, originated predominantly in the dorsal motor nucleus
of the vagus and nucleus ambiguus, are responsible for the parasympathetic autonomic
innervation of most of the thoracic and abdominal viscera along with motor innervation
of the larynx and pharynx, respectively. Its afferent fibers convey visceral information
to the solitary tract nucleus and, sequentially, to the locus coeruleus, hypothalamus,
amygdala, thalamus and insular cortex. However, it is widely known that other brain
regions, such as the spinal trigeminal nucleus, area postrema and reticular formation
of the medulla can receive afferencies as well. Additionally, the vagus nerve is composed
of three types of fibers: myelinated A fibers, predominately responsible for touch
transmission; myelinated B fibers, responsible for visceral stimuli transmission;
and unmyelinated C fibers, responsible for the transmission of pain. The vagus nerve
is comprised mainly of C fibers and its conduction speed is rather slow (8.8 to 12.6
m/s)[18].
In the neck, the vagus nerve lies within the carotid sheath, deep between the carotid
artery and the jugular vein. However, it is important to acknowledge anatomical differences
between the right and left vagus nerves, primarily when planning for a surgical procedure.
The preferential implantation of the electrode on the left is due to the innervation
of the sinoatrial node by the right branches, which poses a greater risk of cardiac
arrhythmias[3].
Mechanisms of action
The exact mechanism through which VNS exerts antiepileptic effects has not been completely
elucidated yet. Although it has been demonstrated that type A fibers are the most
excitable ones, followed by types B and C, respectively, it was once believed that
all fibers should be stimulated to suppress seizures. Subsequently, scientists have
found that C fibers are the ones responsible for the EEG desynchronization associated
with epileptiform activity abolishment. Nevertheless, successive research has demonstrated
that this effect was seen even after lesion of C fibers, suggesting that A and B fibers
probably play a significant role[19].
Nowadays it is well established that VNS influences locus coeruleus and raphe nuclei
to modulate cortical activity through alteration of noradrenergic and serotonergic
projections[3]. The augmentation of locus coeruleus activity after electrical stimulation of the
vagus nerve, demonstrated by an increase in c-fos, may provoke release of noradrenaline
in the limbic circuit and activation of the dorsal raphe nucleus, which send diffuse
serotonergic projections to the diencephalon and telencephalon. It is clear that VNS
therapy induces variations in regional blood flow in different cortical areas including
the thalamus, mesial temporal lobe, prefrontal cortex and limbic circuit, which is
supported by neurofunctional imaging. Indeed, it has been postulated that modulation
of some specific areas, such as the limbic circuit, could be related to better outcomes[20].
Surgical procedure
The surgical technique for VNS implantation was initially described by Reid[21] in 1990 and consists of coiling an electrode around the left vagus nerve and placing
a generator in an infraclavicular pocket, which takes approximately one to two hours.
The procedure starts with a horizontal cervical incision at the level of the cricothyroid
membrane, from the midline to the medial border of the sternocleidomastoid muscle.
After opening the platysma, it is often necessary to divide the omohyoid muscle to
expose the carotid sheath. The vagus nerve is identified deep between the carotid
artery and jugular vein and is later individualized to allow the placement of an electrode
array of three spirals (a tethering coil, an anode and a cathode). The generator is
implanted next, in a subfascial pocket in the left infraclavicular area, after tunneling
the distal end of the electrode subcutaneously and securing the connections. The VNS
is turned on 10 days after the procedure so one can differentiate adverse effects
of stimulation from vagal dysfunction due to surgical manipulation. Besides being
reversible and not causing neuroablation, the device can safely be explanted if needed[22], as in cases of lead breakage/malfunction or infection.
VNS settings
The VNS device allows programming of three fundamental parameters: output current,
frequency and pulse width, in addition to on and off times ([Figure 2]). Although the settings currently used ([Table 1]) were derived initially from animal studies followed by human studies (EOS 1 to
5), they have not been clearly defined. Individual variations are considerably frequent
due to the lack of conclusive randomized research that objectively compares different
parameters. Ideal stimulation should target the delivery of the least amount of energy
that would be sufficient to activate afferent fibers (responsible for the therapeutic
effect) without compromising efferent fibers (responsible for side effects), and still
augment battery life. The conduction velocity of efferent fibers, formed mainly by
A fibers, are higher than the afferent ones, but the rheobase and chronaxie are the
same or slightly different. In addition, the waveform and direction (anode placed
proximally or distally) of stimulation apparently does not influence thresholds. Considering
that A and B fibers are possibly the ones responsible for the VNS effects, understanding
the complexity of their stimulation should enhance our knowledge on how to properly
stimulate the vagus nerve. In fact, if these parameters could be monitored in an implanted
patient, these data could be used as biomarkers to titrate stimulation and optimize
therapy[23].
Figure 2 Diagram of VNS stimulation parameters.
Table 1
Suggestion of parameter adjustments on subsequent appointments.
|
Variable
|
1
|
2
|
3
|
4
|
5
|
6
|
7
|
8
|
|
Output current (mA)
|
0.25
|
0.5
|
0.75
|
1
|
1.25
|
1.5
|
1.5
|
1.5
|
|
Frequency (Hz)
|
20/30
|
20/30
|
20/30
|
20/30
|
20/30
|
20/30
|
20/30
|
20/30
|
|
Pulse width (μs)
|
250/500
|
250/500
|
250/500
|
250/500
|
250/500
|
250/500
|
250/500
|
250/500
|
|
ON time (sec)
|
30
|
30
|
30
|
30
|
30
|
30
|
30
|
30
|
|
OFF time (min)
|
5
|
5
|
5
|
5
|
5
|
5
|
3
|
1.8
|
|
Magnet current (mA)
|
0.5
|
0.75
|
1
|
1.25
|
1.5
|
1.75
|
1.75
|
1.75
|
|
Magnet ON time (sec)
|
60
|
60
|
60
|
60
|
60
|
60
|
60
|
60
|
Courtesy of Cyberonics, Inc.
The output current varies from 0.0 to 3.5 mA, but initial programming is started at
0.25 to 0.5 mA. It is then gradually titrated monthly up to 1.75 to 2 mA, as the majority
of vagal fibers will already be stimulated with currents around 1.5 to 2.25 mA[24]. It has been shown that the outcome in the first three months after implantation
was very similar in groups that used output currents below and above 1 mA. Nevertheless,
in unresponsive patients, the increase in current was associated with better seizure
control, even though the current effect could be partially dependent on the stimulation
period (increased response after longer stimulation periods)[25]. The threshold for vagal nerve stimulation in children is apparently higher than
in adults, which could indicate the need for higher stimulation parameters (current
or pulse width) to obtain similar effects[18].
In turn, frequency is set around 20 to 30 Hz, as values above 50 Hz can irreversibly
damage the nerve[26] and 1 Hz stimulation is not as effective in controlling seizures[9]. In addition, pulse width is adjusted to 250-500 μs ([Table 1]).
The VNS is a cyclic stimulation with an ‘on time’ that usually lasts 30 seconds and
an ‘off time’ of three to five minutes, although these parameters can be programmed
according to the patient’s response ([Table 2]). In a retrospective analysis of stimulation settings in 154 patients (XE5 study),
it was impossible to correlate a better outcome to modifications in current, frequency
or pulse width after three and 12 months of follow-up. However, decreasing the ‘off
time’ to 1.1 min or less in one group provided seizure reduction of 39%, as opposed
to 21% in the group with baseline settings stimulation[27]. Although some researches advise fast stimulation (7 sec on and 30 sec off), no
statistical superiority has been demonstrated yet[4]. Furthermore, increments in stimulation parameters will drain battery life and raise
the need for generator replacement. Computational models have demonstrated that, even
though a smaller number of fibers would be excited when pulse width is reduced from
500 to 250 μs, the required increase in output current, to keep the desired stimulus,
consumes less energy[24]. Lower values of pulse width (250 μs) and frequency (20 Hz) may also be tried, to
reduce adverse stimulation effects, according to the manufacturer’s manual.
Table 2
Duty cycles for various parameters.
|
Duty cycle (% ON time)
|
|
|
|
OFF time (min)
|
ON time (sec)
|
|
|
|
7
|
14
|
21
|
30
|
60
|
|
0.2
|
58*
|
69*
|
76*
|
81*
|
89*
|
|
0.3
|
44
|
56*
|
64*
|
71*
|
82*
|
|
0.5
|
30
|
41
|
49
|
57*
|
71*
|
|
0.8
|
20
|
29
|
36
|
44
|
59*
|
|
1.1
|
15
|
23
|
29
|
35
|
51*
|
|
1.8
|
10
|
15
|
19
|
25
|
38
|
|
3
|
6
|
9
|
12
|
16
|
27
|
|
5
|
4
|
6
|
8
|
10
|
18
|
|
10
|
2
|
3
|
4
|
5
|
10
|
*Not recommended; Courtesy of Cyberonics, INC.
Besides the programmed stimulation provided automatically according to the predefined
settings, the VNS system also allows an independent activation induced by the patient
through a magnet, with the purpose of aborting an evolving seizure. This magnet-induced
stimulation uses an output current of 0.25 mA higher than usual for twice the ‘on’
time. Boon et al.[28] was one of the pioneers in evaluating vagus nerve stimulation efficacy and noticed
seizure interruption after magnet use in almost 60% of patients after three years
of follow-up.
Seizure reduction rate
In 1999, a compilation of five clinical studies examining the long-term efficacy of
VNS was published[29]. Four hundred and forty patients with partial (415 patients) or generalized (25
patients) epilepsy were followed for up to three years (396 for one year, 188 for
two years, and 93 for three years). The response rate was 36.8% at one year, 43.2%
at two years and 42.7% at three years, with 35% reduction in seizures in the first
year, 44.3% after the second and 44.1% after the third year. Another review that included
1,104 implanted patients followed for two years[30], also corroborated the persistent effects of VNS. Subsequently, others have demonstrated
VNS efficacy with increasing rates of responsive patients after four years (69%)[31] and five years (64%)[32], with mean seizure reduction of 76% after 10 years (p < 0,05)[33]. Additionally, approximately 27% of implanted patients monitored for two years remained
seizure free for more than one year[34].
In addition to focal epilepsy, VNS is also effective in treating other types of seizures,
as initially shown by two pilot uncontrolled studies. According to Tecoma et al.[35], in a series of five patients with generalized epilepsy, two were seizure free and
two were responders after six months. In Lennox-Gastaut syndrome there were promising
results[36], with a 58% reduction in seizure frequency six months after implantation in a multicenter
retrospective study of 50 patients[37]. SCN1A gene mutations might improve as well, as reported by Fulton et al.[38], who described a 75% rate of responders in 12 patients.
Similarly, VNS also played a relevant role in the management of refractory epilepsy
in children, as observed in one of the major retrospective multicenter studies, which
analyzed 347 children and found that 32.5% were responsive patients after six months,
37.6% after 12 months and 43.8% after 24 months[8]. When comparing patients younger and older than 12 years, no significant difference
in efficacy or complication rate was demonstrated after five years of follow-up in
141 patients[39], although some reported increased infection rates in children[40].
Predictors of response have not been completely clarified, although some elements
may have a prognostic value. In a cohort of 70 patients with partial or generalized
epilepsy, there was an increase in response rate from 54% to 77% in patients younger
than five years[41], which is in contrast to a multivariate analysis that demonstrated increased rates
of seizure freedom in those who had an epilepsy onset above 12 years of age[42]. Besides age, unilateral EEG epileptiform activity was also correlated with a higher
chance of seizure freedom[43]. In a series of 400 patients, the only predictive factor was focal change in the
EEG (p = 0,004), apart from a tendency toward better results in focal epilepsy (p
= 0.09)[44]. However, when analyzing different seizure types, generalized seizures were significantly
associated with higher rates of seizure freedom during the first year, when compared
to partial seizures[42].
Apart from these encouraging results, some could not confirm VNS efficacy. According
to Hoppe et al., when compared to the best medical treatment, VNS associated with
AEDs apparently did not benefit epileptic adults after one- or two-year follow-ups[45]. According to these authors, the efficacy of medical treatment is underestimated
and VNS side effects may compromise quality of life, which justifies their results.
However, it is essential to consider that VNS candidates often have catastrophic epilepsy,
low functional capability and worse prognosis with progressive deterioration, regardless
of therapy, which could compromise results.
Is of great importance to acknowledge that VNS effects are not immediate and seizure
control improves gradually[46]. Moreover, the therapy may alter the course of the disease and reduce its progression
as well.
[Table 3] shows a summary of various studies evaluating high frequency VNS outcomes.
Table 3
Summary of evidence of high stimulation VNS outcome in the management of refractory
epilepsy.
|
Authors and year
|
Class of evidence
|
No. of patients
|
Follow-up (mos)
|
> 50% reduction (%)
|
Mean or median % reduction (%)
|
|
Ben-Menachem et al., 1994[9]
|
I
|
67
|
3.5
|
38.7
|
30.9
|
|
George et al., 1994[47]
|
II
|
24
|
16–18
|
NR
|
52
|
|
Salinski et al., 1996[11]
|
II
|
100
|
12
|
18,40
|
32
|
|
Handforth et al., 1998[12]
|
I
|
94
|
3
|
NR
|
28
|
|
Ben-Menachem et al., 1999[48]
|
II
|
64
|
20 (3–64)
|
NR
|
45
|
|
Amar et al., 1999[49]
|
I
|
164
|
15
|
39
|
37 and 45*
|
|
Labar et al., 1999[50]
|
II
|
24
|
3
|
45.8
|
46
|
|
Vonck et al., 1999[34]
|
II
|
15
|
29 (12–48)
|
66.6
|
57.1
|
|
Parker et al., 1999[51]
|
III
|
15
|
12
|
26.6
|
17
|
|
Murphy et al., 1999[52]
|
III
|
51
|
18
|
NR
|
42
|
|
DeGiorgio et al., 2000[13]
|
II
|
195
|
12
|
35
|
45
|
|
Kawai et al., 2002[31]
|
III
|
13
|
56 (48–91)
|
69
|
63
|
|
Chavel et al., 2003[53]
|
III
|
23
|
24
|
61
|
54
|
|
Murphy et al., 2003[54]
|
III
|
96
|
32 (12–108)
|
45
|
NR
|
|
Uthman et al., 2004[55]
|
III
|
25
|
6–144
|
60
|
52**
|
|
Saneto et al., 2006[56]
|
III
|
43
|
18 (7–40)
|
51
|
51
|
|
De Herdt et al., 2007[57]
|
III
|
138
|
44 (12–120)
|
59
|
51
|
|
Ghaemi et al., 2010[43]
|
III
|
144
|
36 (24–71)
|
62
|
NR
|
|
Englot et al., 2011[30]
|
III
|
1104
|
24
|
56
|
62
|
|
Elliot et al., 2011[58]
|
III
|
65
|
124
|
NR
|
76.3 and 80*
|
|
Elliot et al., 2011[44]
|
III
|
400
|
59 (3–136)
|
64
|
55.8 and 59.2*
|
|
Ching et al., 2013[33]
|
III
|
100
|
6–144
|
51
|
49
|
|
Yu et al., 2014[59]
|
III
|
69
|
12
|
41
|
40
|
|
Orosz et al., 2014[8]
|
II
|
347
|
12
|
38
|
NR
|
|
Serdaroglu et al., 2016[60]
|
III
|
56
|
87
|
62.5
|
NR
|
|
Pakdaman et al., 2016[61]
|
II
|
44
|
60
|
11
|
67
|
NR: not reported; *mean and median; **evaluation at 144 mos.
Quality of life
In addition to efficacy and seizure reduction, quality of life is another aspect that
should be considered when evaluating VNS therapy, as psychosocial factors also contribute
to score improvements. After assessing the quality of life of 17 VNS patients for
one year, with a questionnaire regarding memory, physical and emotional well-being,
depression and functional limitations (QOLIE-10), Ergene et al.
[62] demonstrated that all scores improved significantly, regardless of seizure reduction
(p < 0.01). Similar results were obtained in a cohort of 136 patients implanted with
VNS, in which not only responsive but also unresponsive patients notably improved
after three months of follow-up (p < 0.0015 and p < 0.005 respectively), with no statistical
difference between the groups[46]. On the other hand, when analyzing quality of life in 19 children with Lennox-Gastaut
syndrome, no statistical improvement in cognitive and behavioral scores could be detected
after two years. Nevertheless, this should be carefully considered, as these children
are usually severely impaired and outcomes could be adversely impacted[36].
Adverse effects
The most common side effects consist of hoarseness, dysphagia and coughing (recurrent
laryngeal nerve stimulation or damage), discomfort or pain in the oropharynx (superior
laryngeal nerve stimulation or damage) and dyspnea. These symptoms are generally induced
by stimulation and may be very frequent during therapy initiation or after settings
adjustments (approximately 60% of patients), but tend to decrease over time. Less
frequent symptoms include bradycardia, asystole and facial paresis. These events are
commonly managed by modifying stimulation parameters, such as reducing pulse width,
without impairing seizure control. In 48 VNS patients, 14 experienced adverse effects
using output currents between 1-3 mA, but improved completely after pulse width reduction
from 500 to 250 or 130 μs, without an increase in seizure frequency[63]. Therefore, VNS therapy is well tolerated, with adverse effects predominantly induced
by stimulation and generally reversible. Irreversible nerve damage, in turn, is usually
rare[44].
Whereas studies report rates of 1% to 5% of hardware malfunction[4], these estimates are highly variable. Révész et al.
[40] published a 3% incidence of lead breakage, noticed mainly by the increase of seizure
frequency, as not all fractures are identified on imaging screening. Infection rates
vary from 3% to 7% and, although generally treated with intravenous antibiotics and
explantation of the device, some have described success managing these patients exclusively
with oral antibiotics.
Cost-effectiveness
Although initial studies had demonstrated that VNS seemed to be an expensive therapy,
long-term evaluations of emergency department visits and intensive care unit admission
costs showed that these exceed VNS expenses during and after the implant. This could
be justified by the increase in battery life after settings adjustments and the progressive
rise in the number of responsive patients[64]. In a retrospective study of 536 adults
who underwent VNS implantation, there was a reduction of 17% in emergency department
admissions in the first year (p = 0.03) and 42% in the second year (p = 0.01).
Future Perspectives
Currently, one of the main goals in therapy improvement is to decrease patient risks,
which can be achieved by increasing battery life and reducing the number of surgical
procedures to replace the generator. According to the manufacturer’s manual, setting
decrements could augment device durability. A considerable percentage of VNS patients
improve with low output currents, even after a belated course, as stimulation effects
are not immediate. Moreover, decreasing frequency from 30 to 20 Hz and pulse width
from 500 to 250 μs does not reduce the number of stimulated fibers and, consequently,
does not interfere in treatment efficacy.
Another approach to improve patient care and increase battery life is the use of a
rechargeable generator. This is already used in a few stimulation devices for the
treatment of Parkinson’s disease, dystonia and pain and allows stimulation to last
for approximately nine years, compared to three years with non-rechargeable batteries.
Although it may significantly reduce expenses, the rechargeable system has some disadvantages,
mainly related to the need for routine charging of the battery. Possible permanent
damaged in cases where it is not charged in an adequate time frame may occur as well.
For epilepsy treatment with vagus nerve stimulation, a rechargeable system designated
ADNS-300 has been tested in three patients. Its generator has a rechargeable battery
that lasts for 12 years and its electrode consists of a spiral cuff, which contains
two stimulation contacts (cathode and anode) and three recording contacts. Although
output current parameters are empirically adjusted, it is possible to change settings
in this design according to the recorded nerve activity, as was performed in two of
the three patients[65].
The development of new electrode models can also contribute to therapy improvement.
A new system used to treat cardiac insufficiency applies trapezoid instead of square
waves to provide unidirectional stimulation through a cuff electrode in order to reduce
side effects by decreasing external current loss. Preliminary analyses in epileptic
patients have shown similar results to the VNS system, without side effects with stimulation
of up to 2 mA[66].
Transcutaneous vagus nerve stimulation is another safe and well-tolerated alternative,
developed to reduce surgical risks. In a randomized study comparing transcutaneous
stimulation with placebo stimulation, there was a statistically significant reduction
in seizures and improvement in quality of life after one year of follow-up[67].
The identification of response predictors would be of major importance in the improvement
of therapy efficacy. Although some have linked EEG patterns, age of epilepsy onset
and seizure types to better outcomes, as described previously, this association has
not been reported consistently by all authors and no definite biomarker has been validated,
decreasing the likelihood of properly selecting a patient population who would benefit
the most. These could be justified by the heterogeneity of published data that makes
comparison between series extremely difficult. Likewise, a thorough insight of the
mechanism of action would promote an enhanced understanding of VNS parameters and
possibly a larger success rate. Higher current intensities and longer pulse widths
have been shown to increase firing of locus coeruleus neurons, which would increase cortical norepinephrine levels and, consequently, reduce
seizure frequency. However, it has been demonstrated that, in some situations, VNS
response is maximal at moderate stimulation intensity, which could be explained by
neurotransmitter depletion or inhibition mechanisms[68]. Therefore, future research
to analyze therapy efficacy in homogeneous populations and to elucidate the areas
involved in stimulation and their role in seizure control should be further encouraged.
Final Remarks
Vagus nerve stimulation is a safe therapy in the management of adult and pediatric
patients with refractory epilepsy who are not candidates for resective surgery. There
is currently level I evidence for its use in focal epilepsy and level II evidence
for other seizure types. However, approximately one quarter of patients do not benefit
from therapy and few achieve seizure freedom. Therefore, further research must be
done to optimize parameters and improve efficacy.
