When conducting vagus nerve stimulation for
drug resistant epilepsy, which intensity of stimulation
(high or low) should we use?
When conducting vagus nerve stimulation for drug-resistant epilepsy, we suggest to use high intensity stimulation rather than low intensity stimulation (GRADE 1C) (strong recommendation, low level of evidence).
• Supplementary note: Adjustment of stimulation conditions should be conducted in the hospital where the electrode implantation was performed or in a hospital/institution where VNS specialist is present.
The efficacy of vagus nerve stimulation is known to depend on the stimulation conditions. The intensity of stimulation should be adjusted while monitoring its therapeutic effect and adverse effects. Therefore, it is necessary to clarify whether high intensity stimulation or low intensity stimulation is superior when conducting VNS.
In addition, as mentioned in CQ 10-1 “Should vagus nerve stimulation therapy be added to drug therapies for drug-resistant temporal lobe epilepsy?”, we have difficulty in performing comparison between real VNS and sham VNS (with no stimulation). Therefore, there is an increase in randomized controlled trials (RCTs) using low intensity stimulation as sham stimulation (placebo stimulation or pseudo-stimulation) to compare with high intensity stimulation.
There is one Cochrane Review1) on a similar clinical question. This review shows that high intensity stimulation has superior therapeutic effect, while treatment withdrawal is rare both when using high and low intensity stimulation.
There were 4 RCTs that examined the efficacy of vagus nerve stimulation therapy for drug-resistant epilepsy2‒5).
For efficacy, the relative risk of seizure frequency ≤ 50% was 1.74 (95% confidence interval 1.14‒2.65) and NNT (number needed to treat: indicating the number of persons needed to treat to achieve the outcome for one person) was 10. For adverse events, low level stimulation was significantly superior in dysphonia and hoarseness (relative risk 2.06, 95% confidence interval 1.34‒3.17) and dyspnea (relative risk 2.43, 95% confidence interval 1.29‒4.57). Treatment withdrawal, cough, and pain did not differ significantly between high level and low level stimulations.
In all the studies collected, the risk of bias was low overall, and the level was not downgraded for all the outcomes. For inconsistency of the results, I2 was 32% for only dysphonia / hoarseness. Since the effect estimate differed between studies, heterogeneity was considered high. Inconsistency was thus considered serious and was downgraded one rank. There was no problem with indirectness, and was judged not serious. As for imprecision, the confidence intervals in many analyses crossed the clinical decision thresholds, and hence was downgraded by one or two ranks. Regarding publication bias, there were only four studies, and therefore was not downgraded. Consequently, the level of evidence for the outcomes was as follows: “moderate” for seizure frequency ≤ 50%, cough, and dyspnea; “low” for treatment withdrawal, dysphonia/ hoarseness, and pain. The overall level of evidence was “low”.
High level stimulation was superior to low level stimulation for the outcome of seizure frequency ≤ 50%. Among the adverse events, dysphonia/hoarseness and dyspnea showed lower rates in low level stimulation, but since there was no significant difference in treatment withdrawal between two groups, there must be few adverse events serious enough to cause treatment withdrawal. According to expert opinion, many adverse events are reversible and can be controlled by adjusting the stimulation current intensity. Taken together, we decided that high level stimulation is probably superior in terms of the balance between benefits and harms.
We concluded that there is probably no significant uncertainty and variability in patient’s values and preferences because high level stimulation is more effective than low level stimulation, and although adverse events are more prevalent in high level stimulation, they are reversible and can be controlled by adjusting the stimulation current.
Adjustment of stimulation intensity can be done by placing the programming wand over the subcutaneously implanted generator; thus resources and costs are negligible. However, reoperation is needed to replace the generator every few years when the battery runs out. Battery consumption is higher for high level stimulation than for low level stimulation. Based on these, it was decided that high level stimulation costs moderately more as compared to low level stimulation.
In the discussions at the panel meeting, high level stimulation was considered superior in efficacy, and adverse effects were acceptable because most of them were presumably at a level that would not cause treatment withdrawal. As for burden and cost, high level stimulation was expected to consume more battery power, requiring more frequent generator exchange. Based on the above arguments, despite considerable adverse events that did not cause treatment withdrawal as well as the increased burden and cost, we finally unanimously recommended using high level stimulation, considering the highly anticipated seizure control effect.
In Japan, the “Guideline on implementation of vagus nerve stimulation therapy for epilepsy”6) was published by the Japan Epilepsy Society in 2012, which states that “In principle, initiate VNS two weeks after implantation. Start with low stimulation intensity and then gradually increase the intensity while monitoring the adverse effects [recommendation grade C]”.
In 2013, the American Academy of Neurology released a guideline update entitled “Vagus nerve stimulation for the treatment of epilepsy”. There is no recommendation for high level or low level stimulation in that guideline. However, it states that whether stimulation at a higher frequency is more likely to reduce seizures than usual stimulation remains unknown.
For adjusting stimulation intensity, we need a system which is capable of managing complications and coping with equipment troubles.
Further research on the optimal intensity of stimulation is needed. In addition, other than stimulus intensity, there is no RCT on supplementary techniques such as magnet stimulation, which will be a future research subject. It is also desirable to elucidate the mechanisms underlying the subgroup with high response and develop evaluation methods to identify these subjects.
Michael 19932), VNS study Group 19953), Handforth 19984), Klinkenberg 20125)
Appendix CQ10-2-01. Flow diagram and search formula for references
Appendix CQ10-2-02. Risk of bias summary
Appendix CQ10-2-03. Risk of bias graph
Appendix CQ10-2-04. Forest plot
Appendix CQ10-2-05. Summary of Findings (SoF) table
Appendix CQ10-2-06. Evidence-to-Decision table
1) Panebianco M, Rigby A, Weston J, et al. Vagus nerve stimulation for partial seizures. Cochrane Database Syst Rev. 2015; (4): CD002896.
2) Michael JE, Wegener K, Barnes. Vagus nerve stimulation for intractable seizures: one year follow-up. J Neurosci Nurs. 1993; 25(6): 362-366.
3) The Vagus Nerve Stimulation Study Group. A randomized controlled trial of chronic vagus nerve stimulation for treatment of medically intractable seizures. The Vagus Nerve Stimulation Study Group. Neurology. 1995; 45(2): 224-230.
4) Handforth A, DeGiorgio CM, Schachter SC, et al. Vagus nerve stimulation therapy for partial-onset seizures: a randomized active-control trial. Neurology. 1998; 51(1): 48-55.
5) Klinkenberg S, Aalbers MW, Vles JS, et al. Vagus nerve stimulation in children with intractable epilepsy: a randomized controlled trial. Dev Med Child Neurol. 2012; 54(9): 855-861.
6) Kawai K, Sugai K, Akamatsu N, et al. Guideline on implementation of vagus nerve stimulation therapy for epilepsy. Tenkan Kenkyu. 2012; 30(1): 68-72 (in Japanese).
7) Morris GL 3rd, Gloss D, Buchhalter J, et al. Evidence-based guideline update: vagus nerve stimulation for the treatment of epilepsy: report of the Guideline Development Subcommittee of the American Academy of Neurology. Neurology. 2013; 81(16): 1453-1459.
Flow diagram and literature search formula
PICO
P: Patients with drug-resistant epilepsy (children as subgroup)
I: Vagus nerve stimulation at high level stimulation
C: Compared with vagus nerve stimulation at low level stimulation
O: Are seizures controlled (25, 50, 75%)?
Is there a decrease in treatment continuation rate?
Is there an increase in dysphonia/hoarseness? / and cough?
Is there an increase in dyspnea?
Is there an increase in pain?
PubMed search: September 28, 2016
#1 Search ((“drug resistant epilepsy” [mesh] OR ((epilepsy OR seizures OR convulsions) AND (intractable OR refractory))))
#2 Search (“vagus nerve stimulation” [mesh] OR (“vagal nerve” AND stimulation) OR (“vagus nerve” AND “electric stimulation therapy”))
#3 Search (randomized controlled trial [pt] OR meta-analysis [pt] OR randomized OR blind OR observation* OR cohort OR “follow-up” OR cross OR case OR series OR prospective OR retrospective OR placebo OR trial)
#4 (#1 AND #2 AND #3)
Cochrane CENTRAL search: September 28, 2016
(epilepsy OR seizures) AND vagus nerve stimulation
CQ10-2. Flow diagram of literature search (modified PRISMA 2009)
Risk of bias summary
Risk of bias graphs
Forest plot
Summary of Findings (SoF) table
Evidence-to-Decision table
Evaluation table of recommendation decision criteria
Recommendation decision table