Effect of medication withdrawal on the interictal epileptiform EEG discharges in presurgical evaluation
Article Outline
Abstract
Medication withdrawal (MW) is an important method of provoking seizures and activating epileptiform EEG activity during the diagnostic work-up of patients evaluated for epilepsy surgery. Previously it was suggested that MW might influence the seizure-type and activate cortical areas otherwise not producing epileptiform discharges, leading to a false localization of the irritative zone. In order to investigate this we reviewed 42 consecutive cases of MW, of 36 patients, during a 3-year period. We compared seizure frequency, seizure-types and the localization of interictal epileptiform discharges before and after MW. Seizure frequency was significantly higher after MW. In the whole group we found an increase in seizure propagation: the proportion of the complex partial seizures and secondarily generalised seizures increased, while the proportion of the simple partial seizures decreased following MW. In one-third of the patients the interictal EEGs after the MW were different from those recorded before the MW. However, in these discordant cases the EEG findings after the MW (and not before the MW) were concordant with the seizure onset zone and the lesional zone. We conclude that MW is an effective and reliable seizure provoking method, and it does not lead to false localization of the irritative zone.
Keywords: EEG, Medication withdrawal, Presurgical evaluation, Seizure
1. Introduction
The localization of the seizure onset zone (ictal EEG) and the irritative zone (interictal epileptiform EEG discharges) is a key element in the lateralization and localization of the epileptogenic zone in the presurgical evaluation of patients considered for epilepsy surgery.1 The clinical aim and advantage of the medication withdrawal (MW) is to activate the interictal activity and to provoke seizures in order to reduce the duration of the long-term EEG monitoring (LTM).2
The possibility that MW activates cortical areas otherwise not producing epileptiform discharges and/or seizures has been raised.1, 3, 4, 5, 6 This would lead to an erroneous localization of the epileptogenic zone.1 However several studies demonstrated that MW does not induce significant changes in the localization of the seizure onset zone.7, 8, 9, 10
The effect of MW on the localization of interictal epileptiform activity received surprisingly little attention. Ludwig and Marsan3 found a “nonspecific” activation of generalised epileptiform discharges in 63% of their cases. To gain further knowledge on this, we analysed data from our patients who had MW during the presurgical evaluation.
2. Methods
We retrospectively reviewed 42 consecutive cases of MW during a 3-year period (2006–2008) of 36 patients (26 female, 10 male), aged between 16 and 67 years (mean: 39 years). Four patients had 2 cases of MW and 1 patient had 3 MWs, separated by more than 6 months. All patients had localization-related epilepsy (31 had temporal, and 5 had extratemporal foci) and they were evaluated for epilepsy surgery.
The procedure of MW was individualised and determined by the patient-specific parameters, in accordance with suggested guidelines.2 Medication was typically withdrawn to 30–50% of the prescribed daily doses of antiepileptic drugs (AED) within 3–7 days before EEG monitoring in most of the cases.
In 7 cases patients were on monotherapy, in 23 cases on bitherapy and in 12 cases the patients were on 3 or more AED. MW was planned for each patient individually based on their treatment history and seizure frequency in order to minimize the risks for generalised seizures.
Seizure frequency (number of seizures/day) was determined for the period before and after the MW for the 42 cases of MW. To assess whether the possible changes in seizure frequency are biased by the patients who underwent several MW-sessions, we calculated for these patients the seizure frequency before and after MW from all the MW cases of these patients. This allowed comparison of the changes in seizure frequency for the 36 patients. Before MW the patients were admitted to our epilepsy centre, where they were continuously monitored by trained personnel for 5 days, and the seizures were recorded on video. On day 6 the patients were admitted to our epilepsy monitoring unit (EMU). After the registration (2–5 days) the patients were further observed in the epilepsy centre.
Seizures were classified as simple partial, complex partial and secondarily generalised seizures.11 To compensate for the possible bias caused by a large increase or decrease in a seizure-type coming from a few patients, we calculated for each patient who had seizures during the observation period both before and after the withdrawal, the proportion of each seizure-type (number of the seizure-type/total number of seizures) separately for period before and after the MW. This individualised proportion of seizure-types compensates for the bias caused by the different seizure frequency in different patients.
Standard EEGs (at least 30-min duration) recorded before the MW included hyperventilation and intermittent photic stimulation. When this did not result in a reliable localization of the epileptiform discharges, repeated EEGs were recorded (up to 4 times) including sleep-EEGs following sleep-deprivation. After the MW long-term video-EEG recordings (LTM) were performed (up to 5 days). EEGs were evaluated by 2 board-certified clinical neurophysiologists (SB and JA). For the patients with discordant interictal EEG findings before and after the MW we compared the localization of the irritative zone with the localization of the seizure onset zone and with the results of neuroimaging (MRI, SPECT, PET).
We compared the frequency of seizures and the individualised proportions of seizure-types before and after the MW using t-test for paired data. The incidence of different seizure-types in the whole group of patients before and after MW was compared by Fisher's exact test (two-tailed). Comparison of the concordance between the localization of the epileptiform discharges and the ictal onset zone, and the neuroimaging data before and after MW was made by Fisher's exact test (two-tailed).
3. Results
3.1. Seizure frequency
For the 42 cases of MW we found a significant increase in the seizure frequency during the period after withdrawal (mean
=2.65 seizure/day) as compared with the period immediately preceding the withdrawal (mean=0.28 seizures/day) (p<0.001). When comparing the seizure-frequencies of the 36 patients (calculating it from all the MW cases for those patients who had several sessions), the results were almost identical (0.27 vs. 3.03; p<0.001).
In 23 cases there was not any seizure recorded during the 5 days preceding the withdrawal. Out of these in 20 cases there was at least 1 seizure recorded after the withdrawal, and 3 patients did have seizure neither before, nor after withdrawal. Only 1 patient who had seizure before withdrawal failed to produce any seizure after withdrawal. Except for this case and the 3 cases with seizures neither before nor after withdrawal, in all the other cases there was an increase in the seizure frequency.
3.2. Seizure-types
In the whole group of patients, following MW we recorded an increase in the proportion of the complex partial seizures (from 31 seizures=53% to 198 seizures=70%) and secondarily generalised seizures (from 3 seizures=5% to 34 seizures=12%), while the proportion of the simple partial seizures decreased from 42% (25 seizures) to 18% (50 seizures) (p<0.01).
In 18 cases we recorded seizures during the observation periods both before and after the withdrawal. The mean values for the individualised proportions of the seizure-types (number of the seizure-type/number of total seizures for each patient) are shown in Table 1. None of these changes reached the level of significance, although there was a trend for the increase in the individualised proportions of secondarily generalised seizures.
Table 1. Mean values for the individualised proportions of the seizure-types before and after medication withdrawal.
| Seizure-type | Proportion (mean) | p | |
|---|---|---|---|
| Before MW | After MW | ||
| Simple partial seizures | 0.14 | 0.16 | 0.35 |
| Complex partial seizures | 0.81 | 0.65 | 0.07 |
| Partial seizures evolving to secondarily generalisation | 0.05 | 0.17 | 0.06 |
3.3. Side-effects/complications
There were 2 cases of NCSE (2 different patients) in the period following the withdrawal. Both could be easily managed with AED and none of them led to the interruption of the LTM.
3.4. EEG before and after the MW
In one-third (12) of the 36 patients the EEG before MW was different from the EEG recorded after MW (Table 2). Out of these 12 patients, the result of the interictal EEG before the MW was concordant with the seizure onset zone in only 1 patient (number 10 in Table 2). None of these 12 patients had concordant data between the interictal EEG before the MW and the neuroimaging data. In 7 of the 12 discordant patients the localization of the epileptiform discharges after MW was concordant with the seizure onset zone, and in 8 patients it was concordant with the neuroimaging data. The interictal EEG recorded after MW correlated better than the EEG recorded before MW, with the seizure onset zone (p=0.02) and the neuroimaging data (p=0.001). In 4 of the 12 discordant patients the EEG recorded before MW was normal. If we excluded these, and we only analysed the 8 patients where the localization of the irritative zone was different before and after MW, we obtained similar results (p=0.02 and 0.007 respectively).
Table 2. Localization of the interictal epileptiform discharges (before and after MW), seizure onset zone and lesional zone for the 12 patients with discordant data; l, left; r, right; bi, bilateral, synchronous; T, temporal; F, frontal; none: abnormalities in none of the regions (i.e. “normal” result).
| Patient | Localization of the epileptiform discharges | Seizure onset zone | Neuroimaging | |
|---|---|---|---|---|
| Before MW | After MW | |||
| 1 | l-T; l-F | r-T | r-T | r-T |
| 2 | None | l-T | l-T | l-T |
| 3 | l-T | r-T | r-T | r-T |
| 4 | None | l-T; r-T | r-T | r-T |
| 5 | l-T | r-T | r-T | r-T |
| 6 | l-T | r-T | Generalised | None |
| 7 | l-T | l-T; r-T | None | r-T |
| 8 | r-T | Multifocal | Multifocal | Multifocal |
| 9 | l-T | r-T | r-T | r-T |
| 10 | None | l-T | None | l-T |
| 11 | None | r-T; l-T | l-T | l-T |
| 12 | l-T | r-T | r-T | r-T |
4. Discussion
MW was proved to be an effective method of provoking seizures and epileptiform EEG activity, during the presurgical work-up of patients with focal epilepsy. In accordance with previous studies, we found a significant increase in seizure frequency following MW.2
In the whole group there was a significant increase in seizure propagation (complex partial seizures and secondarily generalised seizures) as opposed to the decrease in simple partial seizures. This is in agreement with previously published data.6, 10, 12 However, when we compared the individualised proportions of seizure-types, the difference failed to reach the level of significance. This suggests that the increase in the complex partial and secondarily generalised seizures in the whole group of patients originates from a marked increase in a few patients.
Previously it had been suggested that MW activates cortical areas otherwise not producing epileptiform discharges and/or seizures leading to false localization of the epileptogenic zone.1, 3, 4, 5, 6 The only study addressing the changes in the localization of the interictal epileptiform discharges following MW showed a “nonspecific” activation of generalised epileptiform discharges in 63% of their cases.3 We found a difference in the localization of the interictal EEG discharges before and after MW in one-third of our patients. When calculating this, we also counted as discordant those cases where the EEG after the MW showed a clear-cut and localized interictal focus, but the EEG before MW was normal (i.e. the interictal discharges were scored as localized to “none” of the regions). If we excluded the patients where the EEG before the MW was normal, we only found discordant localizations before and after MW in 22% of the patients (8 out of 36).
For the 12 discordant patients we attempted to elucidate in which condition (on-medication or following MW) the localization of the interictal epileptiform discharges is more reliable. Therefore we compared the localization of them with the seizure onset zone and the lesional zone. The localization of the interictal epileptiform discharges after MW showed a significantly better correlation with the seizure onset zone and the lesional zone than those recorded under medication. Our results do not support the concern that MW might cause activation of cortical areas unrelated to the epileptogenic zone, thus leading to false localizations. Contrary, our results suggest that it is rather the medication that distorts the localization of the interictal epileptiform discharges.
Our EEG data could be theoretically biased by the different recording-times before and after MW: repeated (up to 4) standard recordings of 30min duration (including sleep EEG and the classic provocation methods) vs. continuous recording (after MW). However, we did not analyse the quantity of the epileptiform discharges but only their localization. Up to 4 standard EEG recordings (including sleep EEG) increase the sensitivity of the method to up to 90%.13 On the other hand epileptiform discharges will be found in most persons with epilepsy undergoing monitoring during the first hour of spontaneous sleep and if they are not seen in this period, there is little chance of finding them by examining a prolonged recording in its entirety.14
Our results further support that MW is a safe, effective and reliable method for increasing the number of seizures and activating the interictal epileptiform activity, during the presurgical assessment of patients with focal epilepsy.
References
- . Presurgical evaluation of epilepsy. Brain. 2001;124:1683–1700
- . Antiepileptic drug withdrawal in presurgical evaluation: advantages, disadvantages, and guidelines. In: Luders HO editors. Textbook of epilepsy surgery. Informa Healthcare; 2008;p. 580–587
- . EEG changes after withdrawal of medication in epileptic patients. Electroencephalogr Clin Neurophysiol. 1975;39:173–181
- . Ictal effects of anticonvulsant medication withdrawal in epileptic patients. Epilepsia. 1981;22:297–307
- . Falsely localizing ictal onset with depth EEG telemetry during anticonvulsant withdrawal. Epilepsia. 1983;24:344–355
- . Patterns of seizure activation after withdrawal of antiepileptic medication. Neurology. 1985;35:1537–1543
- . Effects of drug withdrawal on location of seizure onset. Epilepsia. 1986;27:423–431
- . Changes in seizure activity following anticonvulsant drug withdrawal. Neurology. 1990;40:407–413
- . Clinical and electrographic effects of acute anticonvulsant withdrawal in epileptic patients. Neurology. 1991;41:508–512
- Influence on ictal seizure semiology of rapid withdrawal of carbamazepine and valproate in monotherapy. Epilepsia. 2002;43:386–393
- Commission on classification and terminology of the ILAE: proposal for revised clinical and electroencephalographic classification of epileptic seizures. Epilepsia. 1981;22:489–501
- Antiepileptic drug withdrawal in patients with temporal lobe epilepsy undergoing presurgical video-EEG monitoring. Epilepsia. 2001;42:251–255
- . Effectiveness of multiple EEGs in supporting the diagnosis of epilepsy: an operational curve. Epilepsia. 1987;28:331–334
- . Modern electroencephalography: its role in epilepsy management. Clin Neurophysiol. 1997;110:1671–1697
PII: S1059-1311(10)00002-6
doi:10.1016/j.seizure.2010.01.001
© 2010 British Epilepsy Association. Published by Elsevier Inc. All rights reserved.
