Vagus Nerve Stimulation for Medically Refractory ... - Springer Link

19 downloads 0 Views 87KB Size Report
son because the aim of epilepsy surgery and VNS are quite different. While resective .... Naritoku D, Willis J, Treig T, Barolat G, Wernicke JF, First. International Vagus ... Naritoku DK, Tecoma ES, Henry TR, Collins SD, Gilmartin. RC, Labar DR ...
Acta Neurochir (Wien) (1999) 141: 447±453

Acta Neurochirurgica > Springer-Verlag 1999 Printed in Austria

Vagus Nerve Stimulation for Medically Refractory Epilepsy; E½cacy and Cost-Bene®t Analysis P. Boon1, K. Vonck1, T. Vandekerckhove2, M. D'have1, L. Nieuwenhuis1, G. Michielsen1, H. Vanbelleghem1, I. Goethals1, J. Caemaert2, L. Calliauw2, and J. De Reuck1 1 Department of Neurology, Epilepsy Program University Hospital, Gent, Belgium 2 Department of Neurosurgery, University Hospital, Gent, Belgium

Summary Introduction. Vagus nerve stimulation is a novel treatment for patients with medically refractory epilepsy, who are not candidates for conventional epilepsy surgery, or who have had such surgery without optimal outcome. To date only studies with relatively short follow-up are available. In these studies e½cacy increased with time and reached a maximum after a period of 6 to 12 months. Implantation of a vagus nerve stimulator requires an important ®nancial investment but a cost-bene®t analysis has not been published. Patients and Methods. Our own experience with VNS in Gent comprises 15 patients with mean age of 29 years (range: 17±44 years) and mean duration of epilepsy of 18 years (range: 4±32 years). All patients underwent a comprehensive presurgical evaluation and were found not to be suitable candidates for resective epilepsy surgery. Mean post-implantation follow-up is 24 months (range: 7±43 months). In patients with follow-up of at least one year, e½cacy of treatment in terms of seizure control and seizure severity was assessed one year before and after the implantation of a vagus nerve stimulator. Epilepsy-related direct medical costs (ERDMC) before and after the implantation were also compared. Results. A mean reduction of seizure frequency from 14 seizures/ month (range: 2±40/month) to 8 seizures/month (range: 0±30/ month) was achieved (Wilcoxon signed rank test n ˆ 14; p ˆ 0:0016). Five patients showed a marked seizure reduction of V50%; 6 became free of complex partial seizures, 3 of whom became entirely seizure free for more than 12 months; 2 patients had a worthwhile reduction of seizure frequency between 30±50%; in 2 patients seizure frequency reduction has remained practically unchanged. Seizure freedom or V50% seizure reduction was achieved within the ®rst 4 months after implantation in 6/11 patients. Before the implantation, the mean yearly epilepsy-related direct medical costs per patient were estimated to be 8830 US$ (n ˆ 13; range: 1879±31129 US$; sd ˆ 7667); the average number of hospital admission days per year was 21 (range: 4±100; sd ˆ 25:7). In the 12 months after implantation, ERDMC had decreased to 4215 US$ (range: 615±11794 US$; sd ˆ 3558) (Wilcoxon signed rank test n ˆ 13; p ˆ 0:018) and the average number of admission days to 8 (range: 0±35) (Wilcoxon signed rank test n ˆ 13; p ˆ 0:023). Conclusion. VNS is an e¨ective treatment of refractory epilepsy and remains e¨ective during long-term follow-up. Cost-bene®t analysis suggests that the cost of VNS is saved within two years following implantation.

Keywords: Vagus nerve stimulation (VNS); refractory epilepsy; EEG; epilepsy surgery; cost-bene®t analysis.

Introduction Current antiepileptic drug (AED) therapy for patients with seizures is estimated to achieve satisfactory control in about 70% of patients. 10 to 20% of newly diagnosed patients with epilepsy prove to be intractable despite optimal medical treatment [11]. Epilepsy surgery is a therapeutic alternative for such patients but requires a thorough patient evaluation [3]. Patients who have a well-de®ned and resectable epileptogenic area can expect a successful outcome after surgery [4]. In our centre only 30 to 40% of patients entering the pre-surgery evaluation programme ultimately have resective surgery [3, 4]. Patients who are not o¨ered surgery may continue to have frequent, disabling seizures which limit their ability to work and participate in activities. Many also su¨er from the chronic e¨ects of long-term, high-does AED polytherapy [9]. Results of acute and longer-term e½cacy of VNS in chronic epilepsy, mainly in patients with partial onset seizures and during a relatively short period of followup, have been published. The ®rst randomised, blind active controlled study, E03, in 114 patients was published in 1995 (17). At the behest of the regulatory authorities in the USA, the study was repeated, E05, in 196 patients and published recently [14]. Both of these studies had a prospective baseline and patients, if they still met the inclusion/exclusion criteria at the end of the baseline, were randomly allocated to receive either high or low stimulation. High stimulation was a ther-

448

P. Boon et al.

Table 1. General Patient Characteristics N

Sex

Age (years)

Seizure duration (years)

F.U. (months)

History

Seizure type before VNS

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

M F M F M M M F F F F F F F M

36 33 38 32 29 20 30 23 32 24 39 44 17 24 20

15 8 13 31 13 19 25 20 10 23 32 25 12 14 4

43 43 37 35 30 30 29 23 20 18 16 12 12 11 7

febrile seizures, head trauma head trauma head trauma, arachnoid cyst encephalitis none febrile seizures head trauma febrile seizures, head trauma febrile seizures febrile seizures, encephalitis forceps birth meningitis, ventricular atrial drain none none meningitis, head trauma

CPS ‡ SG SPS ‡ CPS ‡ SG CPS ‡ SG CPS SPS ‡ CPS ‡ SG CPS ‡ SG CPS ‡ SG SPS ‡ CPS ‡ SG CPS ‡ SG CPS ‡ SG CPS ‡ SG CPS ‡ SG CPS CPS ‡ atonic CPS

apy thought to be bene®cial and low stimulation was thought to be less bene®cial. Both stimulation modes elicited a subjective sensory stimulus, which guaranteed the ``blinding'' of the study. In both studies, patients with partial onset seizures and a mean age in the early 30 s, having had epilepsy for a mean of approximately 22 years, were shown to have a statistically signi®cant seizure reduction following VNS with the high stimulation parameters (30 Hz, 500 ms, on/o¨: 30 sec/5 min; versus low stimulation: 1 Hz, 130 ms, on/ o¨: 30 sec/180 min) for 12 weeks compared with baseline. In a follow-up study e½cacy increased with time being maximal 12 months after implantation [7]. Side e¨ects, safety and tolerability have also been addressed in several studies [8, 12]. The implantation procedure was typically uneventful. Only one serious adverse event occurred, a generator short circuit caused an increase in output current leading to intense stimulation, pain, hoarseness and irreversible paralysis of the left vocal cord. Commonly reported long-term side e¨ects of vagus nerve stimulation were coughing, hoarseness and throat pain. These symptoms were signi®cantly more prevalent, during stimulation in the high VNSmode group. Abdominal pain, nausea, dyspnoea, chest pain, muscle pain and muscle twitches were reported in high VNS-mode group. Decreasing the output improved symptoms. Pulse rate, blood pressure, respiration rate, weight and temperature were not signi®cantly altered. No signi®cant electrocardiographic changes were reported. In E03 a small increase in gastric acid output was found, but after on year there was no clinical evidence of an increased occurrence of peptic ulcer or other dyspeptic disorders. The mean

AED serum levels remained unchanged during VNS when compared to baseline. Only one case of foetal abnormality has been reported and was associated with concomitant AED polytherapy [2]. Patients and Methods At our centre, 15 patients, with medically refractory partial seizures, have been implanted with a vagus nerve stimulator between 1995 and 1998. They were selected from a population of 90 patients who underwent an extensive presurgical evaluation which included scalp video-EEG monitoring, optimum MRI, interictal FDG-PET and neuropsychological assessment [3]. Some patients also underwent invasive video-EEG monitoring with intracranial electrodes. After thorough presurgical evaluation, 65 of 90 patients (72%) were no longer considered for resective surgery. The non-surgical candidates were o¨ered one of three alternatives, continuing drug therapy with a rematching of their standard AED's (n ˆ 31), participation in phase-3 drug trials with novel AED's such as topiramate, gabapentin or levotiracetam (n ˆ 19), or VNS (n ˆ 15). 15 patients gave informed consent for implantation of a vagus nerve stimulator. The mean age of our patients was 29 years (range: 17±44 years); mean duration of epilepsy was 18 years (range: 4±32 years). All patients except one were receiving chronic anticonvulsant polytherapy, the mean number of AED's being 3 (range 1±5). The patient characteristics are shown in Table 1; the results of interictal and ictal EEG studies, optimum MR and interictal ¯uorodeoxyglucose-PET are described in Table 2. The Neurocybernetic Prosthesis (NCPTM ) System (Cyberonics Inc, Houston Texas), which comprises a pulse generator and bipolar helical lead with an integral tether [8, 13, 15] was placed during a surgical procedure under general anaesthesia that requires two incisions. The ®rst, for the pulse generator, was approximately 8 cm long and 2 cm beneath the left clavicle and the other approximately 10 cm long at the anterior border of the left sternocleidomastoid muscle. The connectors of the lead were tunnelled from the neck incision to the upper chest incision leaving the helical electrodes at the neck. Once exposed within the carotid sheath, the vagus nerve was carefully freed for a minimum of 3 cm so that the bipolar helical electrodes, and the integral tether for support, could be carefully

449

Vagus Nerve Stimulation for Medically Refractory Epilepsy Table 2. Results of the Presurgical Evaluation N

Interictal EEG

Ictal EEG

MR

PET

1

muscle artefact

R hippocampal atrophy

R temporal hypometabolism

2

Bilateral independent sharp and spike-waves R posterior temporal theta

R-sided recruitment

R frontal dysplasia

3

R temporal theta

no changes

4

L-sided recruitment

5

Di¨use slowing, R temporaloccipital spike-waves R temporal theta

L temporal arachnoidal cyst; bihippocampal gliosis L > R normal

R frontotemporal hypometabolism normal

no changes

6

Bilateral slowing

R-sided recruitment

7

bilateral recruitment

8

Bilateral slowing, occipital transients of sleep L temporal slowing

9

Normal

R temporal recruitment

10

Bilateral posterior sharp waves

11

R temporal spike-waves

12

14

R frontal temporal central slowing R frontal-temporal spikewaves Bilateral posterior spike-waves

bilateral temporal recruitment R > L L hemisphere recruitment R frontotemporal recruitment bilateral recruitment R>L bilateral recruitment

15

Normal

R frontal recruitment

13

L-sided recruitment

placed around the vagus nerve. The electrode was stabilised using tie downs in a gentle 3 cm bend for strain relief and then passed over the sternocleidomastoid muscle where a loop was placed and ®xed to the fascia, again with a tie down. The connectors were ®tted into the header of the pulse generator and tightened using the torque wrench provided. A simple test of the integrity of the implanted lead and pulse generator combination was performed prior to closure. Programming was facilitated with a radio frequency telemetry wand connected to a standard laptop computer loaded with the NCPTM software. The programmable parameters, together with their ranges, are shown in Table 3. Standard parameters and rapid cycle parameters di¨ered in signal-on time (30 s vs. 7 s) and signal-o¨ time (5 min vs. 14 s). Stimulation was initiated when the patient had fully recovered, usually within 2±4 weeks after surgery. The output current was subsequently increased according to patient tolerance. The patients were also provided with a magnet allowing additional stimulation to be carried out by the patient or carer in case of an aura or a seizure. Patients were followed on an ambulatory basis at regular intervals, usually every 2±4 weeks during ramping up. Afterwards patients were seen every 1 to 3 months for further follow-up. At every clinic visit, we assessed seizure frequency, seizure severity, AED's being taken and their dosage as well as any side e¨ects experienced as a result of the stimulation of the vagus nerve. Seizure frequency during the year before and the follow-up period after the implantation were compared using a Wilcoxon signed rank test.

bilateral frontal areas of low T1 and high T2signal single signal abnormality in semioval center R hippocampal and cerebellar atrophy normal R medial temporal lobe atrophy L posterior hippocampal signal abnormality normal R frontal-temporal porencephaly normal bilateral polymicrogyri in Sylvial ®ssure bilateral gyral abnormality medial frontal R > L

R temporal hypometabolism L frontal-parietal hypomentabolism R temporal hypometabolism R temporal hypometabolism L parietal-temporal hypometabolism R hemispheric hypometabolism bilateral temporal-Parietal hypometabolism bilateral basal frontal hypometabolism L > R R temporal-parietal ahypo-metabolism bilateral anterior temporal hypometabolism R lateral temporal hypometabolism R lateral temporal hypometabolism

To calculate the epilepsy-related direct medical costs (ERDMC) the following algorithm was used. ERDMC consisted of 1. cost of AED's; 2. cost of clinic visits; 3. cost of hospital admission(s) and; 4. cost of laboratory tests. A comparison was made between the mean sum of these costs in the year before and after the implantation in all patients in whom follow-up was 12 months or more (n ˆ 13). Equally, the total number of hospital admission days per year before and after the procedure were also compared. A Wilcoxon signed rank test was used. As the aim of the study was to examine the cost of daily ongoing treatment of epilepsy, the costs of speci®c presurgical diagnostic examinations were not taken into account.

Results If all of our patients, the surgical procedure was uneventful with no operative or post-operative complications. All patients were discharged from the neurosurgical unit within 48 hours of implantation. The mean post-implantation follow-up time was 24 months (range: 7±43 months). The mean output current we used was 2 mA (range: 1.25±3.00 mA). Usually, the occurrence of coughing and/or an unpleasant sensa-

450

P. Boon et al.

Table 3. Stimulation Parameters Available with the NCP TM System Parameter

Units

Range

Standard parameter value

Rapid cycle parameter value

Output current Signal frequency Pulse width Signal on-time Signal o¨-time Lead impedance

milliamperes (mA) hertz (Hz) microseconds (ms) seconds (s) seconds-minutes (s, min) kilo ohms (KW)

0±3.5 mA 1±143 Hz 130±1000 ms 7 s±270 s 14 s±180 min < 1±7 K KW

variable 30 Hz 500 ms 30 s 5 min 3±4 K KW

variable 30 Hz 500 ms 7s 14 s 3±4 K KW

Table 4. Comparison of Seizure Frequency, Severity and Seizure-free Periods Before and After VNS N

Nr of AED's pre-VNS

Nr of AED's post-VNS

Mean seizure frequency/month pre-VNS

Mean seizure frequency/month post-VNS

Secondary generalisation pre-VNS

Secondary generalisation post-VNS

Maximum seizure-free period pre-VNS (days)

Maximum seizure-free period post-VNS (days)

Stimulation output current (mA)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

3 3 3 2 4 5 3 3 3 2 3 1 3 3 2

3 3 3 2 4 3 4 3 3 2 3 1 4 3 2

8 3 4 40 4 3 30 4 16 35 8 2 200 (clusters) 30 4

0 0(**) 0(**) 25 0 1(**) 20 0 1(**) 30 4(**) 0(**) 50 (clusters) 30 1(**)

yes yes yes no yes yes yes yes yes yes yes yes no no no

no no no no no yes yes no yes yes no no no no no

7 10 7 3 2 14 1 14 7 1 3 30 20 3 7

450 150 750 5 7 30 14 357 7 45 7 240 60 3 14

2.25(*) 2.25(*) 1.5 2.5(*) 2.5 2.5(*) 2 1.75 2 3(*) 2.25 1.25 1.5(*) 2.5 2.25

* Patients treated with rapid cycle stimulation after 4 months of unsuccessfull stimulation with standard parameters. ** Patients who achieved seizure freedom or 50% seizure reduction within 4 months after implantation.

tion in the throat was the limiting factor for further increasing the output current. In 6 patients who did not respond after 9 months and after maximum tolerance was achieved, stimulation parameters were changed to rapid cycle. The signal on-o¨ time ratio was then programmed to 7 : 14 sec. During follow-up the patients showed a mean reduction of seizure frequency from 14 seizures/month (range: 2±40/month) to 8 seizures/month (range: 0±30/month) (Wilcoxon signed rank test n ˆ 14; mean ˆ 6; p ˆ 0:0016). The patient who only showed clusters of up to 10 seizures a day (200 seizures/month) has a clear-cut reduction of the total number of seizures/month. This patient was not included in the analysis for statistical reasons because of the clustering of seizures. Five patients showed a marked reduction of V50% of seizures. Six patients became free of their habitual CPS, 3 of whom have been entirely seizure free for more than 12 months. Two patients had a

worthwhile reduction of seizure frequency between 30±50%; in 2 patients seizure frequency reduction was less than 30% or remained unchanged. Seizure freedom or > 50% seizure reduction was achieved within the ®rst 4 months after implantation in 7/11 patients. Six of 11 patients, who had frequent secondary generalised tonic-clonic seizures before implantation, did not experience generalised seizures following implantation of the vagus nerve stimulator. The mean maximum seizure free period changed from 9 days (range 1±30 days) to 161 days (range 5±750 days) (Wilcoxon signed rank test; p ˆ 0:0017). The number and average dose of AED's taken by the patients remained unchanged in 12 patients. In 1 patient, tapering 2 of the AED's reduced polytherapy but in 2 patients new AED's had to be administered. Stimulation output current and comparison of seizure frequency and severity before and after stimulation is shown in Table 4. The reported side e¨ects of VNS in our patients were

Vagus Nerve Stimulation for Medically Refractory Epilepsy

451

Table 5. Epilepsy-Related Direct Medical Costs (ERDMC)

at least 50% in 31% of patients. In one third of the patients, there is no response in seizure frequency reduction. Our experience seems to con®rm this e½cacy rate. The e½cacy of VNS in the treatment of refractory epilepsy appears to compare favourably with novel AED's such lamotrigine, topiramate, gabapentin [5, 6, 16]. However, no direct comparison of the outcome between patients treated with VNS and AED's is currently available. Comparison with results of resective epilepsy surgery has not been published either. One could question the meaningfulness of such a comparison because the aim of epilepsy surgery and VNS are quite di¨erent. While resective epilepsy surgery has a curative purpose in most cases, VNS is rather a palliative treatment for seizures which can otherwise not be adequately controlled. Previous reports have mainly published short-term results; few reports on long-term results have suggested that the e½cacy of VNS increases over time and reaches a maximum at 12 months post implantation [7]. In our series however, the patients who responded did so within the ®rst 4 months after implantation. Comparing the results of VNS with the outcome in patients who choose or were o¨ered di¨erent treatments (rematching of standard AED or phase-3 AED) would provide more insight into the clinical relevance of VNS. Unfortunately these data are not available since many such patients were lost to our follow-up because they returned to their referring physicians for further treatment. The main issue in further establishing the clinical e½cacy of VNS in the long-term is to identify the best responders. It should be determined whether there are speci®c types of epileptic seizures or epileptic syndromes which respond better to VNS. This requires large-scale multicentre studies involving a large number of patients. Such a study commenced in 1997. Another important issue is how di¨erent stimulation parameters in¯uence clinical response. While early reports have described the e½cacy and safety of the potentially advantageous rapid cycle stimulation, more information in larger series is required [1]. Other stimulation strategies with an increase in the signal on-/o¨-time ratio also need be investigated. The e½cacy of VNS in children has been addressed in small populations but more information on e½cacy and safety in larger patient series with longterm follow-up is needed. VNS o¨ers a favourable safety pro®le. Most acute side-e¨ects are related to initial stimulation and resolve spontaneously without the need to stop the stimulation [8, 12]. VNS lacks the chronic side e¨ects of

N

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Hospital admissions (d/y)

Costs pre VNS

Costs post VNS

pre VNS

post VNS

($US)

($US)

39 15 4 10 16 17 10 23 7 4 26 5 100 0 0

4 15 0 0 0 0 27 0 10 0 18 0 35 0 0

14284 5467 3548 3666 11500 7575 4626 8002 2997 1879 10196 9919 31129 2826 595

2726 5955 2274 1059 2988 2942 9730 883 3692 615 7911 2221 11794 ± ±

hoarseness, intermittent loss of voice during stimulation, persistent coughing and unpleasant chest or throat sensation in 6 patients. In 3 patients these side e¨ects required a temporary seizure reduction of the output current; but none requested that the stimulation be turned o¨. Nine patients did not report any side e¨ects at all. Before the implantation, the mean yearly epilepsyrelated direct medical costs per patient were estimated to be 8830 US$ (n ˆ 13; range: 1879±31129 US$; sd ˆ 7667); the average number of hospital admission days per year was 21 (range: 4±100; sd ˆ 25:7). In the 12 months after implantation, ERDMC had decreased to 4215 US$ (range: 615±11794 US$; sd ˆ 3558) (Wilcoxon signed rank test n ˆ 13; p ˆ 0:018) and the average number of admission days to 8 (range: 0±35) (Wilcoxon signed rank test n ˆ 13; p ˆ 0:023) (Table 5). Discussion Until recently, patients with medically refractory epilepsy who were not suitable surgical candidates could only be o¨ered a rematching of AED's that had been previously and unsuccessfully prescribed or they could be included in phase-3 clinical study protocols with novel AED's. In the early 90's vagus nerve stimulation (VNS) had arisen as a novel antiepileptic treatment [15]. Previous controlled studies have demonstrated clinical e½cacy and safety of VNS [1, 17]. This treatment appears to have a moderate e½cacy, reducing seizure frequency

452

AED's. Drowsiness, behavioural and cognitive disturbances and ataxia have not been reported. However, long-term safety (> 3 years) needs to be further established in larger patient series. More data on generator and battery life are required as well as follow-up studies of patients in whom the generator and the stimulation lead have been explanted. The mode of action of VNS is presently unknown. Like in many other anti-epileptic therapies, elucidation of its neural mechanisms lag behind knowledge of its clinical e½cacy. Substantial progress in the understanding about the mechanisms of action has been achieved but this topic is beyond the scope of this paper. To date, no reports of cost-bene®t of VNS are available. We choose to calculate the epilepsy-related direct medical costs using a simple algorithm which re¯ects mainly the costs of daily therapy for refractory epilepsy. Our data show that in the year after implantation, there is a statistically signi®cant reduction in ERDMC. This suggests that the cost of the device (8000 US$) and the operation (2000 US$) can be saved within 2.5 years after implantation, while battery life exceeds 4 years. From the literature and our own experience, the following conclusions can be drawn. VNS is an e¨ective and safe treatment for medically refractory epileptic seizures during the ®rst months after implantation. It appears to be equally e¨ective and safe in the ®rst 2 to 3 years and lacks common side e¨ects of AED's. Cost-bene®t analysis is favourable. However, VNS should be considered a palliative treatment and only be performed after a thorough patient selection, excluding patients who may bene®t from epilepsy surgery. Future research should be aimed at elucidating the basic mechanism of action of VNS and identifying the best clinical responders.

P. Boon et al.

2.

3. 4.

5.

6.

7.

8. 9.

10. 11. 12.

13. 14.

Acknowledgements This study was supported by grants BOZF-01104495 and BOZF011A0996 from the University of Gent and grant 1.5.2366.99 from the Fund for Scienti®c Research-Flanders. The authors acknowledge the patient dedication of the Epilepsy Monitoring Unit and Neurosurgery nursing sta¨ at the University Hospital in Gent.

References 1. Ben-Menachem E, Manon-Espaillat, Ristanovic R, Wilder BJ, Stefan H, Mirza Z, Tarver WB, Wernicke JF and First International Vagus Nerve Stimulation Study Group (1994) Vagus

15. 16.

17.

nerve stimulation for treatment of partial seizures: 1. a controlled study of e¨ect on seizures. Epilepsia 35: 616±626 Ben-Menachem E, Ristanovic R, Murphy J (1997) Gestational outcomes in epilepsy patients receiving vagus nerve stimulation. Proceedings of the 51st Annual Meeting of the American Epilepsy Society (poster) Boon PA, Williamson PD (1989) Presurgical evaluation of patients with partial epilepsy; indications and evaluation techniques for resective surgery. Clin Neurol Neurosurg 91: 3±11 Boon P, De Reuck J, Calliauw L, Hoksbergen I, Thiery E, Caemaert J, Achten E, Decoo D, De Somer A (1994) Clinical and neurophysiological correlations in patients with refractory partial seizures and intracranial structural lesions. Acta Neurochir (Wien) 128: 68±83 Brodie MJ, Richens A, Yuen for the UK Lamotrigine/Carbamazepine Monotherapy Trial Group (1995) Double-blind comparison of lamotrigin and carbamazepine in newly diagnosed epilepsy. Lancet 345: 476±479 Faught E, Wilder BJ, Ramsay RE, Reife RA, Kramer LD et al (1996) Topiramate placebo-controlled dose-ranging trial in refractory partial epilepsy using 200, 400 and 600 mg daily dosages. Neurology 46: 1684±1690 George R, Salinski M, Kuzniecky R, Rosen®eld W, Bergen D, Tarver WB, Werniche JF and First International Vagus Nerve Stimulation Study Group (1994) Vagus nerve stimulation for treatment of partial seizures: 3. long-term follow-up on the ®rst 67 patients exiting a controlled study Epilepsia 35: 637±643 Landy HJ, Ramsay RE, Slater J, Casiano RR, Morgan R (1993) Vagus nerve stimulation for complex partial seizures: surgical technique, safety and e½cacy. J Neurosurg 78: 26±31 Mattson RH, Cramer JA, Collins JF et al (1985) Comparison of carbamzepine, phenobarbital, phenytoin and primidone in partial and secondarily generalized tonic-clonic seizures. N Engl J Med 313: 145±151 Murphy JV, Hornig G, Schallert G (1995) Left vagal nerve stimulation in children with refractory epilepsy. Arch Neurol 52: 886±889 NIH Consensus Development Conference (1990) Surgery for epilepsy. US Department of Health and Human Services, National Institutes of Health. Bethesda, Maryland 8: 1±20 Ramsay RE, Uthman BM, Augustinsson LE, Upton ARM, Naritoku D, Willis J, Treig T, Barolat G, Wernicke JF, First International Vagus Nerve Stimulation Study Group (1994) Vagus nerve stimulation for treatment of partial seizures: 2. safety, side e¨ects, and tolerability. Epilepsia 35: 627±636 Reid SA (1990) Surgical technique for the implantation of the neurocybernetic prosthesis. Epilepsia 31 [Suppl] 2: 38±39 Handforth A, DeGiorgio CM, Schachter SC, Uthman BM, Naritoku DK, Tecoma ES, Henry TR, Collins SD, Gilmartin RC, Labar DR, Morris GL, Salinsky MC, Osorio I, Ristanovic RK, Labiner DM, Jones JC, Murphy JV, Ney GC, Whesless JW (1998) Vagus nerve stimulation therapy for partial-onset seizures: a randomized active-control trial. Neurology 51: 48± 55 Terry R, Tarver B, Zabara J (1990) An implantable netic prosthesis system. Epilepsia 31 [Suppl] 2: 33±37 The U.S. Gabapentin Study Group N 5 (1993) Gabapentin as add-on therapy in refractory parial epilepsy: a double-blind placebo-controlled, parallel group study. Neurology 43: 2292± 2298 The Vagus Nerve Stimulation Study Group (1995) A randomised controlled trial of chronic vagus nerve stimulation for treatment of medically intractable seizures. Neurology 45: 224± 230

453

Vagus Nerve Stimulation for Medically Refractory Epilepsy

Comments The study is accurately performed and the results obtained bring further support to what was reported previously on the possible role of vagus nerve stimulation in the treatment of epileptic patients who cannot bene®t from the classical surgical resection procedures. In spite of the relatively scarce material, the paper validates some interesting ®ndings, such as: the possibility of obtaining not only the reduction, but also the suppression of the seizures, even if in a few cases only; the remarkable e¨ect on secondary generalization; the persistence in time of the e¨ect. It con®rms that the main drawback of the procedure is the lack of knowledge of the factors in¯uencing the outcome. In this regard it might be interesting to draw attention to the fact that 2 of the 4 patients with a presurgical seizure frequency of at least 30 per month did not bene®t at all from the procedure and in the other 2 a reduction of seizure frequency of less than 50% was achieved (see Table 4). Finally, the ®ndings of a positive cost-bene®t balance might have relevant practical implications. G. F. Rossi Vagus nerve stimulation (VNS) is a novel, but still controversial treatment of refractory epilepsy (1±4). Two large multicenter studies have been performed. The problem with the treatment is, that the goal is a lower frequency of seizures, not a cure; a seizure-free patient. This is the ®rst barrier for nonbelievers, does 50% decrease in seizure frequency make a di¨erence? In the ®rst VNS-study (EO3) actually only 50% of the patients wanted to continue the stimulation treatment after the study period. The strong economic connection of the trials to the expensive equipment of the manufacturer is an other problem. Studies concentrating on safety, e½cacy, quality of life and cost/bene®t analysis are therefore most welcome in patients treated with VNS. Fifteen patients from Gent have been chosen from a group of patients investigated for resective surgery and found not eligible for that. One could ask, if the patients 1±3 would have been candidates

for resection with the given information, and patients 4 and 7±9 perhaps bene®ted from subdural electrodes, but that is always a question of opinion ± the inclusion criteria are uniform. Operative technique, stimulation criteria and techniques are well presented, the safety of procedure was good and complications few, 6 patients had hoarseness or loss of voice temporarily. 2 patients were even seizurefree, but the mean follow-up time is not yet very long. Half of the 11 patients with generalized seizures did not have these after stimulation any more. In these patients the cost-e¨ectiveness of the treatment was not yet very convincing. Only direct costs one year before and after the operation have been analyzed, and the average pre/postsurgical costs have been compared. If in the long run the quality of life is improved and the employment or other major ®nancial factors in the daily life could be solved better, the procedure will de®netely ®nd more supporters. Large, controlled and randomized studies with longer follow-up are important to obtain that knowledge. This study is a good start in that direction.

References 1. Ben-Menachem E (1998) Vagus nerve stimulation for treatment of seizures? Yes. Arch Neurol 55: 231±232 2. McLachlan RS (1998) Vagus nerve stimulation for treatment of seizures? Maybe. Arch Neurol 55: 232±233 3. Hachinski V (1998) Vagus nerve stimulation therapy. Arch Neurol 55: 234 4. Schachter SC, Saper CB (1998) Vagus nerve stimulation, a review. Epilepsia 39: 677686 M. Vapalahti Correspondence: Prof. P. Boon, Epilepsy Monitoring Unit, Department of Neurology, University Hospital, 185 De Pintelaan, B-9000 Gent, Belgium.