Page 469 - ILAE_Lectures_2015

at present. It is advisable to consult the device manufacturer if there is any doubt before
performing MRI.

Effect in epilepsy

VNS appears to have an abortive and a prophylactic effect, both acutely and chronically in
epilepsy. It is effective in various animal seizure models. Although the double-blind studies
were in partial epilepsy, it appears to be broad spectrum. It has been implanted in a variety of
syndromes, including idiopathic generalised epilepsy and Lennox-Gastaut syndrome12, with
broadly similar results13.

A Cochrane review (2001) addressed the efficacy of high-level versus low-level stimulation,
the latter as active control11. The review only included the two early short-term randomised
and double-blind trials14,15. The review concluded that ‘results of the overall efficacy analysis
show that VNS stimulation using the high paradigm was significantly better than the low
stimulation’. The overall odds ratio (OR) for 50% responders was 1.93 (1.1, 3.3). Any
beneficial effect of low stimulation would tend to reduce the high stimulation OR compared
to placebo. Although direct comparisons may not be valid, of interest are ORs reported in
meta-analysis of studies of some add-on AEDs. For example ORs reported for 50%
responders relative to placebo were, in order of increasing magnitude: 1.59 (0.91, 2.97) for
remacemide, 2.29 (1.53, 3.43) for gabapentin, 2.32 (1.47, 3.68) for lamotrigine, 2.46 (1.61,
3.79) for zonisamide, 2.51 (1.88, 3.33) for oxcarbazepine, 3.78 (2.62, 5.44) for levetiracetam
and 4.22 (2.80, 6.35) for topiramate16,17,18,19. Thus VNS in early studies appears to have a
short-term effect in refractory patients approaching that of some of the AEDs. It is less
effective, however, in the short or long term, than resective surgery in well-selected cases,
including temporal lobectomy for mesial temporal sclerosis.

More recently there has been an updated Cochrane review (2015) of the role of VNS in focal
epilepsy20. Again the objective was to determine the efficacy of high-level to low-level
stimulation. Four trials were included in the meta-analysis and high-level stimulation was
shown to be 1.5 times more effective than low-level stimulation. Overall VNS was well
tolerated, with the side effect profile very similar to that previously reported. Hoarseness and
dyspnoea were amongst the more common side effects and these were seen more frequently
with higher levels of stimulation20. Similar findings have also been reported elsewhere21,
although some have been more circumspect as to whether VNS does offer significant patient
benefit above best drug therapy22.

Patients are generally aware of stimulation due to a feeling in the throat, so it is not possible
to fully blind patients to whether VNS is on or off. Nevertheless the two double-blind
randomised controlled trials in partial epilepsy were carried out when the device was first
introduced and presumably the patients may have been less informed as to the expected
effects14,15. Later studies involve large numbers of patients but are open label. The mean
reductions of seizure frequencies of 24.5% and 28% observed at three months in the
randomised trials indicate rather modest benefits.

However, by contrast with AEDs, whose benefits tend to reduce with time, open studies
consistently show an increasing effect of VNS with time. For example, at one year a median
seizure reduction of 45% has been shown, with 20% of patients achieving a greater than 75%
reduction1,15,16. Most striking however are the longer-term studies with progressive
improvement in the seizure frequency for more than ten years. For example Kuba et al found
85 out of 90 patients implanted continued VNS at five years with a median seizure reduction
of 56% and with 64% enjoying a 50% or greater reduction23. By ten years Elliott et al found
a 75% reduction in 65 patients13. However periods of seizure freedom greater than one year