Acta Neurochirurgica

, Volume 159, Issue 1, pp 131–136

Effectiveness of vagal nerve stimulation in medication-resistant epilepsy. Comparison between patients with and without medication changes

  • María A. García-Pallero
  • Eduardo García-Navarrete
  • Cristina V. Torres
  • Jesús Pastor
  • Marta Navas
  • R. G. Sola
Original Article - Functional

DOI: 10.1007/s00701-016-3027-6

Cite this article as:
García-Pallero, M.A., García-Navarrete, E., Torres, C.V. et al. Acta Neurochir (2017) 159: 131. doi:10.1007/s00701-016-3027-6

Abstract

Background

Vagal nerve stimulation (VNS) response is not immediate. A progressive decline in seizure frequency is usually found during a period of 12–18 months after implantation. During this time, the patient’s medication is usually modified, which can create doubts about whether their clinical improvement is due to medication changes or to VNS itself. Our goal is to compare two groups of patients treated with VNS, with and without changes in their medication.

Methods

We prospectively analyze 85 patients who were treated with VNS in our hospital between 2005 and 2014. In 43 patients, changes in the antiepileptic drugs (EAD) were not allowed during the postoperative follow-up and they were compared with 42 patients who were left at the option of neurologist make changes in medication. We analyzed the clinical situation at 18 months and compared the two groups.

Results

Overall, 54.1% of patients had a reduction in seizures of 50% or higher (responders). In the group with no changes in medication, responders reached 63%, while in the group in which changes in medication were allowed, 45.2% were responders. Between responders and non-responders, there were no statistical differences in type of epilepsy, frequency, previous surgery, or intensity of stimulation.

Conclusions

We did not find a statistical difference in seizure frequency reduction between patients with or without changes in medication during their follow-up, so changes in medication did not improve the outcome. Furthermore, the absence of changes in AED can help to optimize the parameters of the stimulator in order to improve its effectiveness.

Keywords

Antiepileptic drugs Epilepsy surgery Seizures Vagus nerve 

Introduction

Vagal nerve stimulation (VNS) is used as a palliative treatment in medication-resistant epileptic patients who are poor candidates for resection or in those in whom resection has failed [1, 4, 7, 12, 23]. The majority of publications showed a 50% or greater decrease in seizure in approximately 50% of patients, and a mean seizure reduction of 43% [7]. VNS response is not immediate. A progressive decline in seizure frequency is usually found during a period of 12–18 months [9, 18, 19]; during this time, antiepileptic drugs (AED) are usually modified [3, 6, 8, 12, 23]. It remains uncertain if the progressive improvement seen in patients treated with VNS might be due to a modification in their medical therapy, due to the VNS or a combination of both. Previously, we have reported a 43 medication-resistant epileptic patient series with a 62.8% significant long-term seizure reduction after VNS in a situation of unchanged medical therapy during 18 months [14]; in that study we did not include a control group.

There has been only other study performed in the adult population on unchanged medical therapy during 1 year after the operation [18]; this study reported that to precisely evaluate the net impact of vagal stimulation in operated patients, it is essential to maintain stable doses of the administered drugs.

Our purpose with this study is to compare the long-term effects of VNS, at 18 months of follow-up, on a group of epileptic patients with no changes in antiepileptic drugs and on a group of patients who were left to the option of a neurologist making changes in medication.

Materials and methods

Between January 2005 and September 2014, 85 adult patients with medication-resistant epilepsy were treated with VNS in the University Hospital La Princesa of Madrid (Spain). There were 49 males and 36 females and the mean age was 33.3 (range, 17–70) years. All patients had undergone a complete preoperative epilepsy evaluation protocol [22, 24], which includes 24-channel scalp video electroencephalography (VEEG), 1.5-T magnetic resonance imaging (MRI), interictal single-photon emission computed tomography (SPECT), and neurologic, psychiatric, and neuropsychological assessment. Patients were presented at a presurgical multidisciplinary conference and recommended for VNS insertion. Typical indications were multifocal or generalized epilepsy not amenable to surgical resection and persistent or recurrent seizures following previous epilepsy surgery.

The VNS was implanted and chronic stimulation was activated 15 days after the surgery. The initial parameters were: 0.25 mA; 20 Hz, 250 ms, 30-s stimulus on-time and 5-min stimulus off-time. Patients were evaluated every 2 or 3 weeks, increasing stimulation by 0.25 mA at each visit up to 2.75 mA in some cases. In the postoperative visits, seizure frequency, AED regimen, side effects, and other significant information reported by family members and caregivers, was collected. Patients were also followed by their referring neurologist in all cases, but only in the first 43 patients were they were asked not to modify any dose of medication during the first 18 months of follow-up.

Seizures were classified, according to their frequency, as daily seizures if the patient had seven or more per week, as weekly seizures if the patient had from 1–6 a week, and as monthly seizures if the patient had less than four per month. Responders were defined as having a 50% or greater reduction in seizure frequency with respect to the mean seizure frequency before the implantation of VNS.

Prospectively, in 43 consecutive patients operated on between 2005 and 2009, changes in the AED were not allowed during the postoperative follow-up [14]. In the following 42 operated patients, it was left at the neurologist’s choice to make changes in medication or not. Patients were followed for 18 months. We analyzed demographic and clinical characteristics, stimulation parameters, and changes in seizure rates relative to historical baseline.

The software SPSS 15.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. Measures of central tendency and dispersion are expressed as median and interquartile range (IQR). Sex, epilepsy type, seizure frequency, previous surgery, and responders were tested with χ2 test. Age and stimulation parameters were tested using the Mann–Whitney U test. Normality was evaluated using the Kolmogorov–Smirnov test. The significance level was set at p = 0.05.

This study was approved by the local ethics committee board.

Results

The demographic and clinical data of all 85 patients are summarized in Table 1.
Table 1

Demographic and clinical data for the 85 patients who underwent VNS therapy for medication-resistant epilepsy

Variable

Number (%) or median (IQR)

Sex (male/female)

49 (57.6%) / 36 (42.4%)

Age at implantation

33 (25–37)

Age at epilepsy onset

7 (2–12)

Epilepsy type

 Generalized or multifocal

39 (45.8%)

 Temporal lobe

12 (14.1%)

 Frontal lobe

20 (23.5%)

 Other extratemporal locations

14 (16.4%)

Baseline seizure rate

 Daily

53 (62.3%)

 Weekly

20 (23.5%)

 Monthly

12 (14.1%)

 Prior failed intracranial epilepsy surgery

20 (23.5%)

Response to VNS according to changes in AED therapy

The main goal of our work is to elucidate the role of AED modifications in the outcome after VNS implantation. Table 2 shows a group of 43 patients with no changes in the AED in the follow-up, and a control group of 42 patients in which it was left to the neurologist’s choice the possibility of making changes in patients’ medication if considered appropriate. As we can observer from that table, there were no differences in clinical and demographic characteristics between both groups.
Table 2

Clinical and outcome differences between a group of 43 patients with no changes in the AED, and a group of 42 patients in which was allowed changes in medication

Variable

No changes in medication [number (%) or median (IQR)]

Changes in medication [number (%) or median (IQR)]

p

Patients

43

42

 

Sex (male/female)

21 (48.8%)/22 (51.1%)

28 (66.7%)/14 (33.3%)

0.09a

Age at implantation (years)

34 (25–40)

32.5 (26.5–36)

0.62b

Age at epilepsy onset (years)

7 (2.5–12)

6.5 (1–12)

0.35b

Epilepsy type

  

0.41a

Generalized or multifocal

22 (51.1%)

17 (40.4%)

 

Temporal lobe

5 (11.6%)

7 (16.6%)

 

Frontal lobe

13 (30.2%)

7 (16.6%)

 

Other extratemporal locations

3 (6.9%)

11 (26.1%)

 

Baseline seizure rate

  

0.06a

Daily

34 (79%)

19 (45.2%)

 

Weekly

5 (11.6%)

15 (35.7%)

 

Monthly

4 (9.3%)

8 (19%)

 

Prior failed intracranial epilepsy surgery

12 (27.9%)

8 (19%)

0.67a

Intensity (mA)

2 (1.5–2.5)

2.5 (2–2.5)

0.02b

Responders

27 (63%)

19 (45.2%)

0.10a

Mean seizure reduction

44.7%

46.6%

0.53a

aχ2

bMann–Whitney U test

No changes in AED: In this group, the mean seizure reduction was 46.6% and there were 27 responders (63%); the mean seizure reduction among the responders was 69.8%. Seven patients (16.2%) had a greater than 90% seizure frequency reduction; four of these seven patients had frontal seizures (one of them was seizure-free), one patient had occipital-temporal seizures, and two patients had generalized seizures. Thirty-four patients (79%) had daily seizures and 12 patients (27.9%) had been operated on previously.

Changes in AED: In the group in which neurologists could make changes in medication, 19 patients were responders (45.2%). The mean seizure reduction in this group was 44.7% and five patients (11.9%) had ≥ 90% seizure frequency reduction. Nineteen patients had daily seizures and eight patients had been operated on previously. Finally, the mean seizure reduction among responders was 64.2%.

The difference in percentage of responders between both groups was not statistically significant (p = 0.10).

Response to VNS overall

As a secondary goal, we analyze the response of VNS in all the patients. At 18-month follow-up, 46 patients (54.1%) were responders, ≥75% seizure control was achieved by 19 patients (22.3%), 50–75% seizure control was achieved by 25 patients (29.4%), and two patients were seizure-free. Between responder and non-responders, there were no statistical differences in type of epilepsy, frequency of seizures, previous surgery, and intensity of stimulation (p > 0.05). Clinical features of responders and non-responders are summarized in Table 3.
Table 3

Clinical features of responders and non-responders after VNS therapy at 18-month follow-up

 

Responders [Number (%) or median (IQR)]

Non responders [Number (%) or median (IQR)]

p (χ2)

Patients

46 (54.1%)

39 (45.8%)

 

Sex (male/female)

26 (56.5%)/20 (43.4%)

23 (58.9%)/16 (41%)

0.52

Age at implantation (years)

33 (28–37)

33 (24.5–39)

0.44

Age at epilepsy onset (years)

6.5 (2–12)

7 (1–12)

0.47

Epilepsy type

  

0.55

Generalized or multifocal

19 (41.3%)

20 (51.2%)

 

Temporal lobe

5 (10.8%)

6 (15.3%)

 

Frontal lobe

10 (21.7%)

7 (17.9%)

 

Other extratemporal locations

12 (26%)

6 (15.3%)

 

Baseline seizure rate

  

0.25

Daily

28 (60.8%)

20 (51.2%)

 

Weekly

15 (32.6%)

12 (30.7%)

 

Monthly

3 (6.5%)

7 (17.9%)

 

Prior failed intracranial epilepsy surgery

10 (21.7%)

10 (25.6%)

0.57

Intensity (mA)

2 (1.5–2.5)

2.5 (2–2.5)

0.62

Discussion

Most of the studies about VNS in medication-resistant epileptic patients have allowed modification of AED types and doses after the implantation [6, 8, 12, 20, 23, 25]. Changes in the AED could mask the true effect of the VNS, and could prevent us from making the appropriate changes in the stimulation parameters.

As we previously reported, this is the only long-term prospective study that has been performed on VNS for adults with no changes in their medical treatment during the follow-up [14], so the progressive seizure rate decline we observed must be attributed to VNS alone.

Now, we have compared the previously reported group of epileptic patients with no changes in AED with a control group of patients who were left to the option of a neurologist making changes in medication. Both groups are homogeneous, with no statistical differences in clinical or demographic features except in intensity of stimulation; the median of intensity of stimulation in the group with no changes in medication was 2 mA (range, 1.5–2.5 mA) and in the group with changes in AED it was 2.5 mA (range, 2–2.5 mA); this difference was statistically significant (p = 0.02). We needed less intensity to get similar results in the group with no changes in medication.

We have obtained 63% of responders in the group without changes in the AED and 45.2% of responders in the other group. This difference was not statistically significant. It is possible that a higher sample size might have shown a better outcome in patients where medication changes were not allowed, but larger studies are needed to confirm this hypothesis.

It is possible that the introduction of new AED could interfere with the evaluation of seizure frequency and side effects of VNS that would impede the optimization of the stimulation parameters. It is also possible that some drugs were discontinued to introduce new ones, which would have lead to the patients’ transient worsening of their seizures. Therefore, the absence of medication changes may be helpful to optimize stimulation settings and thus to improve response rates.

Furthermore, Labar in 2004 [18] reported a retrospective study in 269 patients on unchanged AED in 1-year follow-up with 58% of median seizure rate reduction. This study reported that to evaluate the net impact of vagal stimulation in operated patients, it is essential to maintain stable doses of the administered drugs.

Our results are similar to those obtained in institutions with extensive experience in epilepsy and vagal stimulation [3, 7, 11]. They usually allow modification of AED types and doses after the implantation, so changes in medication therapy during the period of adjustment of parameters do not seem to benefit patients.

In all of our patients, the percentage of responders, 18 months after VNS implantation, was 54.1%, and the mean reduction in seizure frequency was 45.5%. The majority of publications showed a 50% or greater decrease in seizure in a mean of 50.9% (range, 18.4–67%) and the mean seizure reduction in these publications was 42.8% (range, 28–66%) [7].

There are several studies that show contradictory results regarding clinical features that could predict responsiveness to VNS therapy. For example, the most appropriate age of implantation appears to remain subject to argument, since a younger age at the initiation of treatment has been considered a positive predictive factor for VNS response in various studies [15, 18], whereas other publications have found no statistical correlation [11]. We have not found a correlation between age of implantation and VNS response (p = 0.22).

Englot et al. [13] considered that seizure freedom was predicted by generalized seizure type, however, a focal epileptogenic focus has been shown to be a positive predictor of success by multiple investigators [11, 15, 17, 25]. In our series, 27 out of 46 patients with focal epilepsy (58.6%) and 19 out of 39 patients with generalized or multifocal epilepsy (48.7%) were responders, but we did not find a significant difference between them (p = 0.55). The small numbers of patients in this cohort could be responsible for that; further study with a larger sample size could be necessary to reach significant differences.

The relationship between the clinical outcome and the intensity of stimulation was also analyzed. There are several studies suggesting that the high intensity of stimulation corresponds to a better outcome [4, 16, 21], although some other authors consider the time of exposure to the treatment the key factor that leads to appropriate responses [5, 9, 17, 18]. In our series, the outcome was uniformly considered at 18 months and the outcome differences according to intensity of stimulation were not statistically significant (p = 0.62). We think that efficacy substantially improves over 1 year because of an accumulative effect of VNS in seizures, however it is necessary to reach a certain intensity to obtain optimal results.

Finally, the existence of a previous surgery has been suggested as a negative predictive factor of response for some publications [2], while other authors say that a prior failed surgery does not affect the seizure outcome [10, 25]. In our series, 20 patients had been previously operated on; 14 of them by extratemporal resection, three by temporal resection, one by subpial transection, and two by callosotomy. This prior surgery did not affect the outcome, since ten out of 20 patients (50%) had satisfactory results from stimulation (p = 0.57).

All of these conflicts in the literature support the need for more studies on VNS for epileptic patients on unchanged medication to elucidate which are the main predictive factors of responsiveness to VNS.

Our study eliminates the confounding factor of medication changes during the assessment period after VNS implantation and it is controlled with patients under VNS and without changes in medication, but studies with a larger number of patients and long follow-up periods are needed to confirm these results.

Finally, one of the limitations of our study is that the group of patients where medication changes were allowed was operated on later than the group without medication changes. However, one could expect that the most modern group could have done better than the group operated on previously, given the possibility of better operating techniques, more surgery experience, or new drugs. However, the group without medication changes had a similar outcome than the control group, with a statistically significant and less need for current intensity.

Conclusions

We did not find a difference in seizure frequency reduction between patients with or without changes in medication during their follow-up after VNS implantation. Compared to other studies in which changes in medication therapy are allowed, we have obtained similar results, so changes in AED during the VNS adjustment period do not seem to benefit patients. Furthermore, the absence of changes in AED can help to optimize the parameters of the stimulator in order to improve its effectiveness.

There are a lot of conflicts in the literature about the main predictive factors of responsiveness to VNS; therefore, more studies on VNS for epileptic patients on unchanged medication are needed.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflicts of interest.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Copyright information

© Springer-Verlag Wien 2016

Authors and Affiliations

  • María A. García-Pallero
    • 1
  • Eduardo García-Navarrete
    • 1
  • Cristina V. Torres
    • 1
  • Jesús Pastor
    • 2
  • Marta Navas
    • 1
  • R. G. Sola
    • 1
  1. 1.Department of NeurosurgeryUniversity Hospital La PrincesaMadridSpain
  2. 2.Department of NeurophysiologyUniversity Hospital La PrincesaMadridSpain

Personalised recommendations