Keywords
vagus nerve stimulation - vagal stimulation - VNS - refractory epilepsy
Palavras-chave
estimulação do nervo vago - estimulação elétrica - nervo vago - epilepsia resistente
à medicamentos
Introduction
One of the techniques used in the treatment of patients with refractory epilepsy that
are not candidates for resective surgery is vagus nerve electrical stimulation (VNS).[1]
[2] This technique consists of implanting a bipolar electrode in the vagus nerve in
the cervical region and a generator in the infraclavicular region. The left side is
the chosen one for this procedure due to the cardiac fibers that originate from the
right vagus nerve. Although the mechanism of action by which it operates has not yet
been fully elucidated, it is believed to involve diffuse effects on brain metabolism
from both cortical and subcortical regions, through modulation of the activity of
the locus ceruleus, nucleus of the solitary tract and reticular formation of the brain stem.[3] Results show that this therapy is effective not only in reducing the frequency and
duration of epileptic seizures,[4]
[5]
[6] but also in promoting a better quality of life for these patients.[7]
Mechanism of Action
It is postulated that the cortical modulation exerted by VNS is manifested through
the modulation of the noradrenergic and serotonergic systems, especially due to stimulation
of the locus ceruleus and the dorsal raphe nuclei, which was confirmed through measurement of monoamines
in cerebrospinal fluid (CSF).[8] It is known that the increase of the activity of the locus ceruleus after electrical stimulation of the vagus nerve, demonstrated by an increase in c-fos,
can cause both release of norepinephrine in the limbic circuit as well as activation
of the dorsal raphe nuclei, which send diffuse serotonergic projections to the telencephalon
and diencephalon.[9]
Anatomy
The vagus (X) nerve is a mixed cranial nerve with ∼ 80% of sensitive fibers. Efferent
fibers innervate the larynx and promote parasympathetic control of the heart, lungs,
and abdominal viscera. It exits the brainstem at the posterolateral sulcus of the
medulla with the glossopharyngeal (IX) and accessory (XI) nerves.[10] The right vagus nerve innervates the sinoatrial node while the left innervates the
atrioventricular node. The ideal nerve location for VNS implantation is the cervical
region, where it travels in the carotid sheath, between the carotid artery and the
jugular vein. A segment of ∼ 3 cm is commonly needed for implantation and, when feasible,
it should be performed as distal as possible in case a new electrode is needed in
the future.
Surgical Procedure
As described earlier, the device is preferentially implanted on the left side of the
patient to avoid the cardiac fibers of the right vagus nerve. The electrode and generator
are tested before the procedure. The first surgeon is at the patient's left in the
cervical region. The patient lies supine on the surgical table with the head supported
by a cushion and slightly extended; a pad is placed under the shoulder for assistance.
A 5-cm longitudinal incision is made at the level of the cricothyroid interval from
the midline to the anterior border of the sternocleidomastoid muscle ([Fig. 1]). The platysma muscle is divided in the direction of the fibers, and the deep cervical
fascia is opened. The sternocleidomastoid muscle is folded laterally to expose the
neurovascular bundle through blunt dissection. The carotid sheath is opened to expose
the carotid artery and the jugular vein, which is retracted laterally to reveal the
vagus nerve trunk deep in between structures ([Fig. 2]). After careful dissection, the lead's spirals are wrapped around the nerve from
the proximal to the distal contact ([Fig. 3]). It is of great importance to maintain the adequate position of spirals (the anchor
tether is placed inferiorly; the positive contact in between; and the negative contact
superiorly, as demonstrated in [Fig. 4]) to stimulate afferent and non-efferent fibers.
Fig. 1 A 5cm longitudinal incision is made at the level of the cricothyroid interval from
the midline to the anterior border of the sternocleidomastoid muscle
Fig. 2 The carotid sheath is opened to expose the carotid artery and the jugular vein, which
is retracted laterally to reveal the vagus nerve trunk deep in between structures
Fig. 3 After careful dissection, the lead's spirals are wrapped around the nerve from the
proximal to the distal contact
Fig. 4 Adequate position of spirals (anchor tether is placed inferiorly, positive contact,
in between, and negative contact, superiorly)
To place the generator on the anterior chest wall inferior to the clavicle, an incision
is made at the level of the anterior axillary line to create a subcutaneous pouch
under the pectoral fascia, large enough to accommodate the device ([Fig. 1]). The electrode is then carefully tunneled from the neck to the chest, above the
sternocleidomastoid muscle and the clavicle, and connected to the generator. The intraoperative
test is performed to confirm adequate system functioning through impedance testing.
If the implanted generator supports closed-loop stimulation, it is also necessary
to verify the system's capability of correctly identifying the heart rate. At our
center, stimulation is initiated immediately after the procedure with the following
settings: 0.25 to 0.5 mA, 30 Hz, 500 micros, 30 seconds ON, and 5 minutes OFF.
Rationale
The device employed for the electrical stimulation of the vagus nerve allows adjustment
of several parameters, as current, frequency, pulse width, ON and OFF time. Although
the current stimulation protocol was initially based on animal studies[11]
[12]
[13] and, subsequently, on humans (mainly in EOS 1–5 studies), it has not yet been thoroughly
elucidated and individual variations are quite frequent,[14] mainly due to the lack of conclusive randomized trials objectively comparing different
values of frequency, amplitude, and pulse width. Therefore, the present study intends
to demonstrate the safety of reducing the stimulation frequency from 30 to 20 Hz.
Methods
The current study consisted in a double-blind prospective analysis of patients with
refractory epilepsy previously submitted to VNS implantation who underwent reduction
of frequency stimulation (from 30–20 Hz) and were followed up for evaluation of changes
in frequency and/or duration of epileptic seizures and emergence of side effects.
The eligibility criteria included individuals of both genders from 2 to 18 years of
age, with refractory epilepsy of focal or generalized origin, already submitted to
VNS implantation at Hospital Pequeno Príncipe by the same surgeon (T. O.), that demonstrated
interest in participating voluntarily. The exclusion criteria, in turn, consisted
of age group outside the previously mentioned range or lack of interest in participating
in the research. The project was approved by the ethics committee of Hospital Pequeno
Príncipe and did not generated expenses for the participants. All individuals who
agreed to participate in the survey signed the informed consent. It was clarified
that there could be a reduction in the number of surgical procedures for generator
replacement due to increase in battery survival. All changes in stimulation parameters
were performed in the hospital, and the patient remained in place long enough for
at least two cycles of stimulation to occur to early diagnose any immediate side effects.
It should be noted that the researcher responsible for setting adjustments was not
the same who evaluated the results. The participants and their families were extensively
instructed on the double-blind nature of the project and the need for randomization
to reduce placebo effect.
Six patients stimulated with 30 Hz were initially followed up for 3 months with a
questionnaire and a seizure diary. Due to lack of participant compliance, the authors
opted for collecting data through online diaries, telephone contacts or office visits.
After baseline evaluation, the frequency stimulation was reduced to 20 Hz, and the
patients were followed up for another month. Because of the small sample size, adjustments
were performed for all patients, which composed the control group. However, to maintain
the double-blind approach, the participants, and the author responsible for evaluating
the results were unaware of this information ([Fig. 5]).
Fig. 5 Methods' flowchart
In addition to a comparative statistical analysis performed using Microsoft Excel
(Microsoft Corp., Redmond, WA, USA), the data were processed through the t-paired analysis in the Minitab program (Minitab, LLC., State College, PA, USA) to
compare the two stimulation groups (20 and 30 Hz). To test the normality of the distribution
of the 30 Hz and 20 Hz samples according to the multiple variables, the Anderson-Darling
and Ryan-Joiner were used. The H0 hypothesis, rejected if p < 0.05 for 95% confidence intervals, considered that the sample distribution followed
a normal distribution.
Results
Of the 14 patients selected for the research, only 6 demonstrated interest in participating
and signed the informed consent. The mean age of the sample analyzed was 10 years
(8–18 years) and half of the participants was female. In the baseline evaluation,
two patients had already presented with complete seizure remission. An average of
109 seizures per week per patient was observed, with 90.7% being partial, ∼ 9% drop-attack,
and 0.3% being tonic-clonic. All patients who were still seizing obtained some reduction
in seizure duration, and 75% were considered responsive (achieved more than 50% reduction
in seizure frequency). Likewise, three of four patients reported less intense events,
while one remained unchanged.
With the inclusion of all patients in the group of 20 Hz, however, it was not possible
to exclude the possibility that the data did not follow a normal distribution. This
probably occurred due to the presence of two outliers that no longer had seizures
since the baseline evaluation and stayed that way after reducing the frequency to
20 Hz. To approach normality, it was decided to exclude these two participants and
perform all the analysis with only the data of the four remaining ones.
Two patients reported side effects after frequency reduction to 20 Hz: one had transient
dysphonia while the second evolved with permanent dysphagia and dysphonia. Nonetheless,
it is crucial to mention the latter had clinical deterioration due to hospitalization,
which could have contributed to the complaints.
Contrarily to expectations, in the month following settings adjustment, there was
a 68% reduction in seizures (p = 0.054). Moreover, in the 75% of patients who were still seizing, a reduction in
both intensity and duration of the episodes was noticed. It is necessary to emphasize
that, in the two patients who had no seizures, there was no clinical worsening after
reducing the frequency to 20 Hz.
After comparing the mean of the total number of seizures during stimulation at 30 Hz
and at 20 Hz, there was no statistically significant difference (p = 0.054), despite a tendency of seizure reduction with 20 Hz stimulation. This fact
can be explained by the presence of overlap data in the histogram of the two distributions,
even though the means of total seizure number with stimulation at 30 Hz and at 20 Hz
were distinct ([Graph 1]). When applying the paired t-test, it was not possible to state with 95% confidence that the averages of the total
seizure number with stimulation at 30 Hz or 20 Hz were different since the confidence
interval included 0 (- 4.0; 227.5), as demonstrated in [Table 1]. However, when changing the interval (26.1; 197.4), it became evident that the averages
differed with 90% certainty ([Table 2]). This should be carefully interpreted, however, as the resultant seizure reduction
could be simply a consequence of the stimulation of the vagus nerve itself and could
appear months after the initiation of therapy.
Graph 1 Histogram of the total number of seizures during stimulation at 30Hz and at 20Hz
Table 1
Paired t-test of the total number of seizures with stimulation at 30 and 20 Hz
|
N
|
Mean
|
Standard deviation
|
Standard error
|
|
30 Hz
|
4
|
164
|
70.1
|
35
|
|
20 Hz
|
4
|
52.3
|
39.3
|
19.6
|
|
Difference
|
4
|
111.8
|
72.7
|
36.4
|
95% confidence interval for the mean difference: (- 4.0; 227.5).
T-test of difference from mean = 0 (versus no = 0); t-value = 3.07; p-value = 0.054.
Table 2
Paired t-test of the total number of seizures with stimulation at 30 and 20 Hz
|
N
|
Mean
|
Standard deviation
|
Standard error
|
|
30 Hz
|
4
|
164
|
70.1
|
35
|
|
20 Hz
|
4
|
52.3
|
39.3
|
19.6
|
|
Difference
|
4
|
111.8
|
72.7
|
36.4
|
90% confidence interval for the mean difference: (26.1; 197.4).
T-test of difference from mean = 0 (versus no = 0); t-value = 3.07; p-value = 0.054.
Discussion
The VNS system allows for changes in almost all settings, and most of the patients
are stimulated with the standard protocol of 30 Hz, 500 μs, 30 seconds ON, 5 minutes
OFF, and amplitudes that vary from 0.25 to 2.25 mA.
Frequency values range from 20 to 30 Hz, because it has been demonstrated that frequencies
greater than 50 Hz could cause irreversible nerve damage.[15] A recent study in rats, however, suggested that frequencies between 130 Hz to 180 Hz
as recommended in brain, spinal cord, and trigeminal stimulation, could lead to greater
seizure attenuation than 30 Hz stimulation.[16] Nonetheless, these results have not yet been demonstrated in humans. Low frequency
electrical stimulation (1 Hz), in turn, has not been as effective as high frequency
(30 Hz) in seizure control.[17]
Diversely, pulse width ranges from 250 to 500 μs, and current amplitude from 0.0 up
to 3.5 mA. The amplitude chosen for initial stimulation, however, varies from 0.25
to 0.5 mA and is gradually increased to 1.75/2 mA in months. Most patients do not
benefit from further increases as the vast majority of fibers are already stimulated
with values close to 1.5 to 2.25 mA.[18] Although it has been shown that clinical response with a reduction in epileptic
seizures in the first 3 months after implantation was quite similar in groups that
used current amplitudes lower or higher than 1 mA,[19] it should be noted that in the non-responsive group there was greater improvement
after increasing the amplitude. This fact could be explained not only by settings
adjustment, but also by the stimulation time itself,[9] since clinical response may be delayed on account of the cumulative effect of stimulation.
Furthermore, as children tolerate increases in amplitude better than adults because
of fewer side effects, a need for higher current or pulse width values in these patients
may be noticed.[20]
It is noteworthy that stimulation does not occur continuously, but rather in cycles.
Although the initial stimulation is generally started with cycles of 30 seconds ON
and 5 minutes OFF, the system allows for cycles of 7 to 60 seconds ON and 0.2 to 10 minutes
OFF. In a retrospective analysis of the parameters of electrical stimulation in 154
patients in the XE5 study, it was not possible to correlate seizure control with changes
in current amplitude, frequency, and pulse width between 3 and 12 months of follow-up.
In one group, however, it was observed that reducing the OFF time to ≤ 1.1 minute
led to better control with the reduction in seizures being improved from 21 to 39%.[21] Although some authors advocate rapid electrical stimulation (7 seconds ON and 30 seconds
OFF), it has not yet been possible to demonstrate statistical difference with this
protocol.[4]
Moreover, it should be considered that increases in electrical stimulation parameters
will generate higher battery consumption and a consequent reduction in battery life,[18] in addition to increases in surgical procedures to replace the generator. For example,
computational models have already demonstrated that, although there are less stimulated
fibers with pulse width reduction from 500 to 250 μs, the required increase in amplitude
to maintain the same electrical stimulation consumes less energy than with 500 μs
pulse width stimulation and lower amplitudes.[18] Lower values of pulse width (250 μs) and frequency (20 Hz) can be used on patients
according to the manufacturer's manual (Cyberonics, 2015, VNS Therapy®, Cyberonics
Inc. Houston TX, USA), with the main objective of reducing side effects. A projection
of battery life according to the various current values, pulse width, and frequency
can be seen in [Fig. 6].
Fig. 6 Projection of battery life according to the various current values, pulse width,
and frequency
The reduction from 30 to 20 Hz in the present study showed a reduction of 68% in the
incidence of seizures (p = 0.054) as well as low incidence of permanent side effects (only 1 out of 6 patients).
However, it is necessary to interpret these data with caution since most of the participants
were the ones who responded to therapy. Moreover, this improvement could also result
from the time patients had been treated, as vagus nerve electrical stimulation response
may not be immediate, and its effectiveness may progressively increase over time.
Limitations
The main limitations of our study are the small sample size and the use of the same
patient as his own control group. The next step in the investigation of vagus nerve
electrical stimulation at 20 Hz frequency would be the development of a randomized
study, with a group of stimulation at 20 Hz and a control group at 30 Hz with crossover
after a follow-up period.
Conclusions
When considering the significant reduction in the frequency of epileptic seizures
and the improvement in the quality of life of implanted patients, along with the low
incidence of irreversible or debilitating side effects, it is possible to recognize
that electrical stimulation of the vagus nerve is a safe therapy in the treatment
of pediatric and adult patients with refractory epilepsy who are not candidates for
resective surgery. Although this was a pilot study with a small sample size, it demonstrated
that, in the short term, it is apparently safe to reduce the stimulation frequency
to 20 Hz without compromising the effectiveness of therapy. The subsequent increase
in battery lifetime would, consequently, reduce the need of surgical replacements
of the generator.
