Clinical Neurophysiology xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Clinical Neurophysiology

journal homepage: www.elsevier .com/locate /c l inph

Etomidate activates epileptic high frequency oscillations

1388-2457/$36.00 � 2013 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.http://dx.doi.org/10.1016/j.clinph.2013.07.006

⇑ Corresponding author. Address: Department of Neurology, Epilepsy Center,University Hospital Erlangen, Schwabachanlage 6, 91054 Erlangen, Germany. Tel.:+49 9131 85 36921; fax: +49 9131 85 36469.

E-mail addresses: stefan.rampp@uk-erlangen.de, stefan.rampp@uk-halle.de(S. Rampp).

Please cite this article in press as: Rampp S et al. Etomidate activates epileptic high frequency oscillations. Clin Neurophysiol (2013), http://dx.d10.1016/j.clinph.2013.07.006

S. Rampp a,⇑, H.J. Schmitt b, M. Heers c, M. Schönherr a, F.C. Schmitt d, R. Hopfengärtner a, H. Stefan a,e

a Epilepsy Center, Department of Neurology, University Hospital Erlangen, Erlangen, Germanyb Department of Anesthesiology, University Hospital Erlangen, Erlangen, Germanyc Ruhr-Epileptology/Department of Neurology, University Hospital Knappschaftskrankenhaus Bochum, Bochum, Germanyd Department of Neurology, University Hospital Magdeburg, Magdeburg, Germanye Interdisciplinary Epilepsy Center, Neurological Clinic, University Hospital Giessen and Marburg, Germany

a r t i c l e i n f o h i g h l i g h t s

Article history:Accepted 6 July 2013Available online xxxx

Keywords:EpilepsyActivationEtomidatePresurgical workupHigh frequency oscillations

� Etomidate increases the rate of interictal epileptic activity, including spikes and high frequencyoscillations.

� Etomidate does not induce additional epileptic activity distinct from spikes or HFO seen withoutetomidate.

� Low frequency oscillatory activity triggered by etomidate does not show an association with spike andHFO distribution.

a b s t r a c t

Objective: The short acting anesthetic etomidate has been shown to provoke epileptic spikes and rarelyseizures. Influence of etomidate on the occurrence of epileptic HFO (high frequency oscillations) howeveris unknown. An HFO inducing effect of etomidate would allow further validation of the substance as aprovocation measure in presurgical evaluation as well as provide insights into the common mechanismsof HFO, spike and seizure generation.Methods: We retrospectively analyzed EEG data from four patients who underwent etomidate activationduring invasive video-EEG monitoring with subdural strip electrodes. Spikes were manually selected inraw data, HFO in band pass filtered data (80–250 Hz). Rate and spatial distribution of HFO and spikesin three segments were compared: immediately after etomidate administration, as well as during slowwave sleep and while awake.Results: Rates of HFO and spikes increased significantly after etomidate administration: Overall averagerates of spikes were 9.7/min during sleep, 10/min while awake and 61.4/min after etomidate. AverageHFO rates were 9.5/min during sleep, 8.3/min while awake and 24.4/min after etomidate (p < 0.001,non-parametric ANOVA). Spatial distributions of HFO and spikes after administration of etomidate wereconsistent with the seizure onset zone (SOZ) and area of resection when available (SOZ: two patients;resection: one patient; no information: one patient). Except for spurious events, no additional HFO andspike foci were seen with activation.Conclusions: Etomidate administration activates spikes and HFO. Spatial distributions do not extendbeyond electrodes showing spikes and HFO without Etomidate and seem consistent with the epilepticnetwork.Significance: Etomidate activation is a safe procedure to provoke not only epileptic spikes but also HFO,which were shown to have a high specificity for the SOZ.� 2013 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights

reserved.

1. Introduction

Epilepsy surgery requires exact evaluation of a patient’sindividual condition. Crucial to the viability of surgical therapy isthe localization of the epileptogenic network in relation to essen-tial functional areas (Cascino 2004; Rosenow and Lüders 2001). A

oi.org/

2 S. Rampp et al. / Clinical Neurophysiology xxx (2013) xxx–xxx

spectrum of methods is applied to gather this information: Whileimaging may demonstrate structural correlates, only neurophysio-logical investigations like electro- or magnetoencephalography(EEG/MEG) and invasive EEG (iEEG) are able to prove pathophysi-ological involvement and yield information about the functionalorigin of epileptic activity. Functional investigations are especiallyvaluable to evaluate propagation and connectivity within an epi-leptic network (Stefan and Lopes da Silva 2013).

Seizure activity, spikes and sharp waves are the electrophysio-logical targets of such investigations and constitute the markersof epileptic networks in clinical routine. In recent years, high fre-quency oscillations (HFO) have been recognized as an additional,highly specific marker of epileptogenic areas (Staba et al. 2002;Bragin et al. 1999). While the classification of HFO is still being dis-cussed, there is evidence that oscillations between 250–500 Hz(‘‘fast ripple’’), 80–250 Hz (‘‘ripples’’) and 60–100 Hz (‘‘high gam-ma’’) represent distinct entities of fast activity with different de-grees of association with epileptogenic networks: Sloweroscillations demonstrate a more widespread spatial distributionwith a less clear but present concentration in epileptogenic areascompared to fast ripples (Jacobs et al. 2008; Rampp et al. 2010;Worrell et al. 2004). A practical downside of HFO is, that theycan only be reliably recorded using invasive EEG, although recentstudies show preliminary evidence that non-invasive methodsmay be able to detect HFO (Xiang et al. 2009a; Xiang et al.2009b; Rampp et al. 2010; Andrade-Valenca et al. 2011).

In patients with rare epileptic events, presurgical workup maybe ineffective and possibly beneficial surgery is delayed. Longerrecording durations to record sufficient activity are viable only toa limited degree due to the ensuing burden for the patient, as wellas for financial and logistical reasons. For intraoperative electrocor-ticography during epilepsy surgery, considerably longer recordingdurations are no option at all. Typical procedures to provoke epi-leptic activity are tapering of anti-epileptic drugs (AED) and sleepdeprivation. Furthermore, pharmacological agents can be adminis-tered to increase the quantity of epileptic activity and probabilityof seizures. Clonidine (Kettenmann et al. 2005) and dexmedetom-idine (Mason et al. 2009) induce spike and sharp-wave activity,possibly by post-synaptic a1- and a2-receptor mediated attenua-

Table 1Clinical data.

Patient 1 2

Sex Female MaleAge 28 42MRI Normal Cyst and gliosis left fro

(DD: posttraumatic/poinflammatoric defect, cdysplasia, gangliogliom

Interictal surface EEG Temporo-mesial right and temporalneocortical

Frontal close to midlin

Ictal surface EEG Unclear, bilateral seizure onset UnclearInterictal invasive EEG Temporo-basal/mesial right and

temporopolar leftDiffuse (left and rightbilateral frontal)

Ictal invasive EEG Unclear Bifrontal onset, enhancof the left side

Ictal SPECT Right hippocampus UnclearFDG-PET Right temporal (polar and mesial) –Neuro-psychology Left temporal Left frontal

Surgery – –

Histology – –

Please cite this article in press as: Rampp S et al. Etomidate activates epilepti10.1016/j.clinph.2013.07.006

tion of central noradrenergic transmission and depletion of centralnorepinephrin reduction of central norepinephrin lowers seizurethresholds (Mason et al. 2009), which may consequently also resultin increased interictal epileptic activity. Barbiturates and deriva-tives are GABAA-receptor agonists and thus mainly enhanceGABA-inhibition. However, it has been shown, that e.g. thiopentaland methohexital also lower excitation thresholds and increaseepileptic activity. This effect is clinically utilized for intraoperativeelectrocorticography during epilepsy surgery (Wyler et al. 1987).Next to barbiturates, opioids, e.g. remifentanil, may be used to pro-voke epileptic activity in this context (Wass et al. 2001). Amongother effects, remifentanil inhibits GABAergic interneurons, whichinhibit excitatory neurons under normal circumstances, e.g. neu-rons involved in the generation of epileptic activity. This disinhibi-tion of excitation may consequently result in an increase ofepileptic activity (Siggins et al. 1986).

Intravenous administration of low-dose etomidate has beensuggested as a further alternative activation approach. Few earlyreports have suggested that etomidate may provoke epilepticactivity (Ebrahim et al. 1986; Gancher et al. 1984; Hsieh et al.1990), but were not followed by larger studies. Pastor et al. haverecently demonstrated in a case report that etomidate may triggerseizures, which they applied for ictal single photon emissiontomography (SPECT) (Pastor et al. 2008). In a larger study, theyshowed that the substance may be safely used to induce interictalspiking activity, which they observed in all 20 investigated pa-tients. In contrast, induced seizures were seen in only two patients(Pastor et al., 2010). The activated spikes, recorded by scalp EEGand invasive foramen ovale electrodes correctly lateralized the sideof seizure onset in all but one patient.

In contrast, influence of etomidate on the occurrence of epilep-tic HFO is unknown. An HFO inducing effect of etomidate would al-low further validation of the substance as a provocation measure,e.g. in patients with rare spontaneous activity. In situations withtime constraints, e.g. in studies using MEG for non-invasive detec-tion of HFO or intraoperatively for HFO-tailoring of resection ex-tents (Wu et al., 2010), the ability to trigger HFOs would bebeneficial. Furthermore, etomidate is used as an anesthetizingagent in the etomidate speech and memory test (‘‘eSAM’’) as an

3 4

Male Female52 22

nto-polarst-orticala)

Hippocampal sclerosis right Hippocampal sclerosis left

e, left frontal Fronto-temporo-mesial right, fewspikes left temporal

–

Unclear Left-fronto-temporalfrontal, Temporo-basal and mesial right

(mesial became apparent afteraddition of foramen ovale electrodes),few spikes left temporal

Left temporo-mesial and -basal

ement Temporo-mesial right Left temporo-mesial

––Right temporal/parietal, lefttemporo-lateral

Left temporal

Hippocampal resection and anterior2/3 of the temporal lobe on the rightside. Several seizures in the first3 months after surgery, then seizurefree for 2 years up to now (Engel 2B,Wieser 1)

–

Hippocampal sclerosis –

c high frequency oscillations. Clin Neurophysiol (2013), http://dx.doi.org/

S. Rampp et al. / Clinical Neurophysiology xxx (2013) xxx–xxx 3

alternative to the intracarotid amobarbital procedure or Wada-test(Jones-Gotman et al., 2005). A spike and HFO inducing effect couldenable using invasive EEG during eSAM for focus localization.

We retrospectively analyzed iEEG data from four patients whounderwent etomidate provocation during invasive video-EEGmonitoring. We hypothesized that etomidate increases the rateof ripple events, as well as the rate of spikes. We compared thetopography of spontaneous and provoked ripple oscillations andspikes with the reported seizure onset zone as well as the MRIfindings.

2. Methods

2.1. Participants

We investigated four patients who underwent long-term inva-sive video-EEG monitoring during presurgical evaluation for epi-lepsy surgery: 2 women and 2 men, aged 28, 42, 52 and 22.Three patients suffered from pharmacoresistant temporal lobe,one from frontal lobe epilepsy (patient 2). Table 1 summarizes clin-ical details. The decision to administer etomidate for provocationof epileptic activity, i.e. spikes and seizures, was based on the clin-ical complexity of each case due to rare epileptic activity, unclearfindings or epileptic activity in the dominant hemisphere andwas not influenced by preconditions of the presented study. Specif-ically, triggering of HFO was not a goal of the procedure but wasanalyzed retrospectively. Due to the complex clinical situations,only one patient (#3) could proceed to epilepsy surgery and there-fore usage of outcome as a validation parameter was not feasible.The number of patients included in this study was also limiteddue to the constraints of the clinical decision to perform etomidateprovocation during invasive video-EEG monitoring. All patientsgave their informed consent to undergo the procedure.

2.2. Data acquisition

Patients were investigated during long-term invasive video-EEGmonitoring using subdural strip electrodes with platinum contacts(AD-Tech Medical Instrument Corporation, Racine, WI, USA). Acqui-sition was performed on the third day after implantation. Numberand location of electrodes were determined based on previousfindings of presurgical evaluation. All electrodes were recordedreferenced to a surface electrode at the vertex. Patients weremonitored using ECG, pulse oxymetry and non-invasive blood pres-sure measurements. Etomidate was administered intravenously(0.1 mg/kg body weight) by an anesthesiologist (author HS).

Data for analysis was selected from longer video-EEG monitor-ing recordings. A first 5-min segment starting at etomidate admin-istration was selected. For comparison, two additional 30-minsegments were used. One segment with wake state data was ana-lyzed, taken from the recordings directly preceding the Etomidateinvestigation, except for a pause of 1 h in patient 4 (due to trans-portation of the patient to the MEG lab and setup of the recording(see below)). A second comparison segment was selected fromslow wave sleep from nights also preceding the etomidateinvestigation.

Data were recorded using an IT-Med EEG amplifier using ananalogue band pass filter from 0.08 to approximately 360 Hz. Datawere digitally sampled at 1024 Hz. Data after etomidate adminis-tration in Patient 4 was acquired using the integrated EEG ampli-fier of a Magnes II MEG system (4D Neuroimaging, San Diego,USA) for recording of simultaneous invasive EEG and MEG withan analogue band pass filter from 1 to approximately 120 Hz anda sample rate of 1041 Hz. Especially the low pass filter could notbe set to higher edge frequencies due to technical limitations.

Please cite this article in press as: Rampp S et al. Etomidate activates epilepti10.1016/j.clinph.2013.07.006

However, this was not considered a critical limitation, as rippleactivity was repeatedly found in frequency ranges below 120 Hz.Data from patient 4, however separate from the presented dataand without Etomidate administration has been published (Ramppet al., 2010). MEG data of etomidate administration recordings wasnot used in the present study.

2.3. Analysis

All data segments were visually inspected for spikes (referencemontage vs. surface electrode as the vertex). Artifacts were visuallyidentified and excluded from further analysis. Individual spikeswere visually identified and marked in the channel that showedthe clearest pattern with the highest overall amplitude. For HFOdetection, all data segments were band pass filtered to the ripplefrequency band (80–250 Hz, limited to 80–120 Hz using the 4DNeuroimaging system). Faster oscillations were not considereddue to the limited bandwidth of the amplifiers, which was con-strained to <360 Hz in the IT-Med and to <120 Hz in the 4D Neuro-imaging system. HFO events were also visually identified andmarked in the channel that showed the clearest and earliest pat-tern; however oscillatory patterns were only accepted when theyconsisted of at least 4 complete cycles exceeding the baselinenoise.

Subsequently, rates of spikes and HFO events were determinedfor all data segments, as well as topographic distribution. In addi-tion to comparisons of control, pre- and post-administration seg-ments, topography was also compared to location of reportedlesions, as well as the seizure onset (SO). For visualization of thetime course as well as statistical comparison of the different seg-ments, rates of spikes and HFO were calculated for consecutive10 s intervals.

2.4. Statistical analysis

Spike and HFO rates over time (10 s windows) were submittedto a 2-way analysis of variance (ANOVA) with state (awake, sleep,etomidate) and patient as factors. Data showed a non-normal dis-tribution, demonstrated by a significant Lilliefors test for normal-ity. Although ANOVA is thought to be robust in cases of violatedassumption of normality (Glass et al. 1972; Harwell et al. 1992;Lix et al. 1996), we complemented the analysis by a non-paramet-ric 2-way ANOVA, the Scheirer-Ray-Hare (SRH) test (Dytham 2011;Scheirer et al. 1976). Results were qualitatively similar, we there-fore report the results of the SRH-test. Post-hoc tests were per-formed using the Wilcoxon rank sum test and the bonferroniprocedure to correct for multiple comparison (Matlab R2011b,The MathWorks, Natick, MA, USA). Corrected p-values are re-ported. Time course of event rates instead of e.g. only average val-ues were used to better reflect the different behaviors, with highand low activity intervals over time, as well as to enable a more de-tailed statistical analysis.

Topography of spike and HFO events during sleep and wakestate and after etomidate administration were compared byassessing contacts with more than 10% of events. Contacts withless than 10% were interpreted as displaying only spurious events.Results were then qualitatively compared in regard to overlap un-der the different conditions.

3. Results

3.1. General observations

Approximately 5–10 s after the end of etomidate injection, allpatients lost consciousness for 2–4 min. Symptoms during this time

c high frequency oscillations. Clin Neurophysiol (2013), http://dx.doi.org/

Fig. 1. Example EEG data after etomidate administration (patient 2). The upper traces show raw EEG with spikes and HALFO. The lower traces show the same data segmentfiltered between 80 and 250 Hz to appreciate spike associated and spike independent HFO.

4 S. Rampp et al. / Clinical Neurophysiology xxx (2013) xxx–xxx

were limited to dilated pupils, skewed deviation of the eyes, staringand subtle smiling facial expression. Myoclonia, typically associ-ated with etomidate (Vanlersberghe and Camu 2008), occurredonly in patient 2, who presented with light myoclonia in the upper

Please cite this article in press as: Rampp S et al. Etomidate activates epilepti10.1016/j.clinph.2013.07.006

extremities. No clinically detectable seizures occurred. Further-more, with the exception of staring, no semiological symptoms typ-ical for the habitual patient’s seizures were observed. Moreover,iEEG recordings did not show any signs of rhythmic seizure activity.

c high frequency oscillations. Clin Neurophysiol (2013), http://dx.doi.org/

Table 2Spike and HFO rates, percentage of HFO superimposed on spikes.

Patient Sleep Wake Etomidate

Spike rate (per min)

1 1.4 0.1 41.02 31.1 23.3 63.63 2.7 12.5 85.24 3.3 4.3 55.8HFO rate (per min)

1 2.0 6.7 8.62 32.4 19.2 54.23 2.9 0.6 14.64 0.6 6.6 20HFO superimposed on spikes

1 2% 0% 56%2 17% 4% 51%3 8% 41% 78%4 11% 2% 97%

Table 3SRH-results.

SS SS/Mstotal d.f. p-value

SpikesState (factor) 71400137.3 351.9 2 <0.001Patient (factor) 36569091.9 180.2 3 <0.001State � patient (interaction) 37238216.3 183.5 6 <0.001HFO

State (factor) 23357528.7 230.2 2 <0.001Patient (factor) 12668892.8 187.3 3 <0.001State � patient (interaction) 4558358.1 134.8 6 <0.001

SS = sum of squares; MS = mean square; d.f. = degrees of freedom.

S. Rampp et al. / Clinical Neurophysiology xxx (2013) xxx–xxx 5

However, roughly coinciding with the loss of consciousness and thebeginning of increased spike and HFO activity, high amplitude lowfrequency oscillations (HALFO) appeared, comparable to previousreports (Stefan et al. 2010; Gancher et al. 1984). HALFO distributionwas diffuse and not related to spikes and HFO topography. Fig. 1shows an example of EEG traces after Etomidate.

3.2. Rate of epileptic activity events

The increase of both HFO and spikes started promptly approxi-mately 10–20 s after etomidate injection, coinciding with the lossof consciousness and appearance of HALFO and lasted for approx-imately 5 min. A maximum of HFO activity was observed in the ini-tial part after administration in patients 1 and 4, showing a latermaximum in patient 2 and an approximately even distribution inpatient 3 (Fig. 3). Spike activity peaked in the initial phase afteradministration in patients 1, 2 and 4 and remained at an almostconstant level in patient 3 (Fig. 3).

In comparison to wake and sleep control data, etomidate in-duced a statistically significant increase of HFO and spike rates.Overall average rates of spikes were 9.7/min during sleep,10/min while awake and 61.4/min after etomidate. Average HFOrates were 9.5/min during sleep, 8.3/min while awake and 24.4/min after etomidate (Table 2). The amount of increase variedbetween patients (SRH-test, p < 0.001 for factors and interactions,Table 3; p < 0.001 Wilcoxon rank-sum post hoc test, Fig. 2, Table 2).Spike rates over time were significantly higher during wake versussleep state (p < 0.001, Wilcoxon rank-sum test) and reached high-est values after etomidate administration (p < 0.001 compared toeither sleep or wake state). HFO rates were slightly higher duringsleep than during wake state; however this difference did notreach statistical significance (p > 0.5). Again, highest rates were ob-served with etomidate (p < 0.001). HFO were overlaid on spikes in0–97%, depending on patient and state. In general, percentage washighest under etomidate (Table 2).

3.3. Topography

Spike topography after etomidate administration was similar towake and sleep states. Specifically, no new spike foci were seen. In

Fig. 2. Spike and HFO rates of sleep, wake and post-administration data segments.Boxes denote first and third quartile, center line the median, whiskers the range ofthe data excluding outliers (crosses). Statistical significance of differences in posthoc tests is denoted as brackets at the top of the figure, all shown significantdifferences had p < 0.001.

Please cite this article in press as: Rampp S et al. Etomidate activates epilepti10.1016/j.clinph.2013.07.006

patient 1, more than 10% of spikes were seen in contacts TB2A andTP1A during sleep, which was comparable to the awake state:TB2A and the close more anterior-lateral contact TB1C. Spikes afteretomidate occurred mainly in TB2A. Subclinical ictal activity onlywas observed in temporo-polar contacts on the right side.

In patient 2, left frontal lateral contact FRL1B showed promi-nent spike activity during sleep. Spikes in the awake state howeverdisplayed a more diffuse distribution over fronto-basal contactsleft (FRB1B/C) and right (FRB2D), as well as frontal lateral rightcontacts (FRL2C), After etomidate, spikes were seen in both groupsof contacts: FRB1B/C, FRB2D and FL1B. Seizure activity was dif-fusely distributed over left sided contacts, however in some sei-zures also consisted of bifrontal spiking.

Main spiking activity under all conditions in patient 3 was seenin right temporo-mesial contact TB2A. Independent of the pre-sented results, an additional foramen ovale electrode was im-planted to improve the sensitivity for hippocampal activity. Thiselectrode clearly showed the seizure onset in areas next to the stripcontacts presenting with the maximal spike and HFO rate. Spikeand HFO rates for the foramen ovale electrode were not deter-mined due to the lack of data under etomidate activation. Resec-tion also included the temporo-mesial contacts and resulted inseizure freedom after few initial seizures (Engel 2B, Wieser 1,2 years follow-up, Table 1).

Finally, in patient 4, spikes were seen in temporo-basal to me-sial contacts in all conditions (TB1A/B during sleep, TB1A-C inthe wake state, TA1A and TB1A-C after etomidate), which dis-played the seizure onset during monitoring.

HFO distribution after etomidate was also similar to controlconditions. In patient 1, HFO were detected in temporo-polar con-tacts on the left with only minor differences of single contacts inthe mesial-lateral direction (TP1C during sleep, TP1B in the wake

c high frequency oscillations. Clin Neurophysiol (2013), http://dx.doi.org/

Fig. 3. Time course of spike HFO occurrence rate. Each bar corresponds to the absolute number of detected HFO events in a 10 s window centered at the specific time point.Time scale was readjusted to designate the start of the respective segment (sleep, wake and after etomiate administration).

6 S. Rampp et al. / Clinical Neurophysiology xxx (2013) xxx–xxx

state, TP1A after etomidate), concordant with subclinical ictalactivity during monitoring.

Comparable to the spike distribution, HFO were scattered overseveral contacts in patient 2: left frontal anterior and lateral(FRA1A and FRL1C) during sleep, also left frontal anterior and lat-eral (FRA1A, FRL1B/C) while awake. Etomidate induced HFO in leftfrontal anterior and lateral contacts (FRA1A, FRL1C/E), but also inright fronto-basal contacts (FRB2B/D). The latter show minor HFOactivity (610%) during sleep and while awake.

HFO in patient 3 were focal during sleep (TB2A) and after etom-idate (also TB2A), however more scattered during awake state:temporo-polar (TP2A) and bifrontal (FR2D, FR1B). TB2A and TP2Awere included in the resection.

Patient 4 displayed HFO in left temporo-mesial (TB1A, TC1A)and temporo-lateral (TA1H) contacts during sleep. While awake,HFO occurred also in mesial contacts. HFO activity was constrainedto temporo-mesial areas after etomidate (TB1A/B), concordantwith the seizure onset.

4. Discussion

In the presented study, we investigated influence of etomidateactivation on HFO rate using invasive EEG. While the number ofpatients was small due to the rare circumstances allowing suchan investigation, the results provide significant findings: etomidateinduced a significant increase of both HFO and spikes in contactsthat also showed epileptic activity without administration ofetomidate, in line with the seizure onset zone and resection, whenthis was available. Etomidate capability to provoke spike activityhas been demonstrated before (Pastor et al. 2010; Stefan et al.2010), however exclusively using non-invasive or limited invasivemethods, such as foramen ovale electrodes in combination withscalp EEG. A direct comparison to the gold standard of invasiveEEG has not been performed before. Our results now support andreproduce these previous findings and underline the fact that theactivated spikes and HFO do not occur in new, possibly unrelatedfoci, which is demonstrated by the similar topographies with andwithout Etomidate.

Both HFO and spikes after etomidate showed a large overlapwith areas displaying ictal activity and the seizure onset during vi-deo-EEG monitoring. Furthermore, such areas with HFO and spikeswith and without etomidate were resected in patient 3. While hesuffered few seizures in the first three postoperative months, he re-mained seizure free thereafter (postoperative follow-up of 2 years).Thus, provoked activity was concordant not only with spontaneousactivity, but also with markers of the epileptogenic zone, i.e.seizure onset and good surgical outcome. These findings imply a

Please cite this article in press as: Rampp S et al. Etomidate activates epilepti10.1016/j.clinph.2013.07.006

significant clinical potential; however, validity and reproducibilityremains to be determined in a larger cohort.

In contrast, HALFO were unrelated to the epileptic network;epileptic activity and HALFO were found in disjoint groups of con-tacts. The occurrence of such high amplitude delta oscillations is afrequent observation after etomidate administration (Meinck et al.1980; Gancher et al. 1984; Arden et al. 1986; Pastor et al. 2010)-and are also encountered in non-epileptic patients (Stefan et al.2010).Pastor et al. (Pastor et al. 2010) linked the occurrence ofdelta oscillations under etomidate to its exclusive action onGABAA-receptors and the role of the GABA-system for non-rapideye movement (REM) delta/slow wave sleep (SWS).Hypothetically, etomidate may increase epileptic activity by induc-tion of SWS, which has been shown to facilitate the occurrence ofHFO and spikes, suggesting an at least partially common mecha-nism (Bagshaw et al. 2009). However, the rates of both spikesand HFO after etomidate significantly exceeded rates during sleepcontrol data. Furthermore, while HFO occurred slightly more fre-quent during sleep, spike rates were higher in the wake state. Thusthe effect of induced sleep does not seem sufficient to explain theobservations, but may contribute to a smaller degree especially tomore frequent HFO.

Alternatively, the increase of epileptic activity after etomidatemay be due to disinhibition of excitatory neurons, similar to the ac-tion of remifentanil (Wass et al. 2001): In therapeutic levels, as usedin our study, etomidate does not directly activate GABAA-receptors,but sensitizes them to GABA. The same amount of GABA thus re-sults in strengthened inhibition. Hypothetically, this may impactexcitatory neurons and inhibitory interneurons to a different de-gree. If the effect on interneurons dominates (possibly locally orlimited to specific networks), excitatory target cells are disinhibitedand increase firing rates. This would then correspond to morespikes and HFO in cells of the epileptic network. Since etomidateacts by sensitization of GABAA-receptors, GABAergic drugs or sub-stances increasing GABA-levels, such as anti-epileptic benzodiaze-pines and barbiturates may putatively enhance this pro-epilepticeffect by increasing GABA-levels. Our data is not sufficient to pro-vide supportive or disconfirming evidence; however, it would beinteresting to evaluate this hypothesis in future studies.

The mechanism of action of benzodiazepines, e.g. diazepam,also consists of enhancing the effect of GABA on GABAA-receptors,however, impact on the EEG and specifically on epileptic activity iscompletely different from etomidate. Diazepam induces a generalincrease of beta activity. Based on the idea that GABAA-receptorsmay be reduced in epileptic lesions, local decreased induction ofbeta activity has been utilized for epileptic focus localization (Clauset al. 2009; Gotman et al. 1982). Benzodiazepines are anticonvul-sant and also significantly decrease interictal spiking (Jawad

c high frequency oscillations. Clin Neurophysiol (2013), http://dx.doi.org/

S. Rampp et al. / Clinical Neurophysiology xxx (2013) xxx–xxx 7

et al. 1986); there is no evidence for induction of fast oscillationsabove the beta and gamma band. Next to the opposite impact onepileptic activity, etomidate administration is not followed by anincrease of beta activity (Ghoneim and Yamada 1977). Thus,although both substance groups act on the same receptor, mecha-nisms of action differ significantly, possibly due to differences on asystem level, i.e. proportion of affected receptors in different net-works or network subcomponents.

Gigli and Gotman (Gigli and Gotman 1991) demonstrated thatspiking rates in amygdala kindled cats are primarily influencedby seizure occurrence, i.e. more and stronger seizures are followedby an increased number of spikes. Carbamazepine, although effec-tive to reduce seizures, also increased spikes. Similar effects inhumans are shown by an earlier study in 44 patients with pharma-coresistant epilepsy (Gotman and Marciani 1985). Here, seizuresresulted in an increase of spike rates both during wakefulnessand sleep, which could last for days. Spiking was not particularlyhigh or low before seizures and did not influence occurrence of sei-zures. Interestingly, drug levels of a range of anticonvulsants,including carbamazepine, had no detectable influence on spiking.In contrast, the anticonvulsant topiramate, similar to diazepam,was shown to have a strong inhibitory effect also on spikes (Placidiet al. 2004). Epileptic HFO were shown to correlate with diseaseactivity. Their number increases with medication reduction, how-ever stays stable after seizures, although the generating area, cor-responding to the number of involved channels, increases(Zijlmans et al. 2009). In line with these findings, propofol has dif-ferent effects on spikes, seizures and HFO. It has an antiepileptic ef-fect and reduces the number of epileptic HFO, however does notchange spike rates (Zijlmans et al. 2012). These findings show thatspikes, seizures and HFO reflect different aspects of epileptic pro-cesses. While HFO response to AEDs and diagnostic value are clo-sely related to that of seizures, spiking is more variable and mayrepresent a secondary effect of ictal activity. Reminiscent of Ras-mussen’s ‘‘red’’ and ‘‘green spikes’’, this variability may also stemfrom putative subtypes of spike patterns, with different generatorsand thus specificity for epileptogenic tissue (Engel et al. 2009). Inview of these findings, etomidate may act on mechanisms trigger-ing seizures (Pastor et al. 2008; Pastor et al. 2010) and HFO. Thesignificant spiking increase (Ebrahim et al. 1986; Gancher et al.1984; Hsieh et al. 1990; Pastor et al. 2008; Pastor et al. 2010)may then follow indirectly, similar to elevated spike rates after sei-zures (Gigli and Gotman 1991; Gotman et al. 1982). However, ourdata does not rule out direct impact on spiking or induction of anincreased number of putative ‘‘red’’ spikes, which may partiallyshare common generating mechanisms with HFO and seizures.

Interindividual differences of HFO and spike rate increases andtheir topographic distributions may hypothetically also be causedby inherent characteristics of the generating areas, e.g. temporo-mesial networks may be more susceptible to produce epilepticHFO (Jacobs et al. 2009) and may thus be more accessible to Etom-idate activation. Further studies have to evaluate these hypotheses,especially as the highest total amount of HFO were seen in frontalareas (patient 2), whereas the largest relative effect of Etomidatewas seen in a patient with temporal lobe epilepsy (patient 4,Fig. 3).

In a previous study (Stefan et al. 2010), we reported that thesensitivity for etomidate activated epileptic discharges may behigher in MEG than in surface EEG. In a subsequent study, weadministered etomidate in the setting of Video-EEG-monitoringand thus using optimal conditions for EEG. However, we did not re-cord spikes under etomidate. We hypothesized that superpositionof HALFO may hinder the detection of spikes. Our present findingsnow show a clear increase of spikes under etomidate with similartopography. The electrode coverage in our study was not sufficientto investigate whether this spatial distribution of etomidate

Please cite this article in press as: Rampp S et al. Etomidate activates epilepti10.1016/j.clinph.2013.07.006

induced epileptic activity is more focal and limited in extent com-pared to control data. If this would be the case, the underlyingsmaller neuronal populations would only generate small total cur-rents and thus low amplitudes on the surface. Low spike ampli-tudes within HALFO would therefore mean a particularly reducedsignal to noise ratio. Since sensitivity for small neuronal generatorsin superficial areas is higher for MEG compared to EEG (Goldenholzet al. 2009), detection of spikes under etomidate may therefore beeasier using MEG, at least under specific conditions. CombinediEEG/MEG-recordings or iEEG with more extensive electrode cov-erage may provide more data on the actual extent of activated net-works under etomidate.

In our study, etomidate induced interictal activity but no sei-zures were observed. Patients lost consciousness and showed onlysymptoms that differed from their typical seizure semiology. Inaddition, no side effects except light myoclonia occurred. This cor-responds to previously reported findings. For instance Pastor et al.reported a low probability of 3–9% for a seizure after etomidateadministration (Pastor et al. 2010). Therefore, etomidate shouldbe considered an option to provoke interictal activity for investiga-tional settings with limited recording time, such as MEG.

The main limitations of our study are low patient numbers andlimited coverage by invasive EEG. The former was caused due tothe constraints of the clinical decision to perform etomidate prov-ocation during invasive video-EEG monitoring. Retrospective rein-vestigation of EEG recordings or future prospective studies duringeSAM procedures (Jones-Gotman et al. 2005) could provide addi-tional information about the value of etomidate to induce epilepticactivity, especially as this aspect may interfere with speech andmemory testing during eSAM. Topographic features of spikes andHFO after etomidate will have to be reinvestigated in patients withmore extensive coverage, e.g. also during eSAM or intraoperatively.Further studies should additionally evaluate the impact on fast rip-ple oscillations, which were shown to provide the highest specific-ity for the epileptogenic zone (Bragin et al. 1999). Limitations ofthe amplifiers used in the presented work prevented us to investi-gate this.

In summary, etomidate induced a significant increase of HFOand spikes rates. Spatial distribution of the induced activity didnot go beyond previously active areas in conditions without etom-idate. Moreover, distributions were consistent with the zone of sei-zure onset or early propagation when this information wasavailable. Etomidate may thus be a useful option for activation ofmainly interictal activity in invasive EEG and MEG or for intraoper-ative ECoG.

Acknowledgements

This study was supported by the ELAN fund of the UniversityErlangen (09.05.30.1) and the Deutsche Forschungsgemeinschaft(DFG, STE 380/15-1).

References

Andrade-Valenca LP, Dubeau F, Mari F, Zelmann R, Gotman J. Interictal scalp fastoscillations as a marker of the seizure onset zone. Neurology 2011;77:524–31.

Arden JR, Holley FO, Stanski DR. Increased sensitivity to etomidate in the elderly:initial distribution versus altered brain response. Anesthesiology1986;65:19–27.

Bagshaw AP, Jacobs J, LeVan P, Dubeau F, Gotman J. Effect of sleep stage on interictalhigh-frequency oscillations recorded from depth macroelectrodes in patientswith focal epilepsy. Epilepsia 2009;50:617–28.

Bragin A, Engel J, Wilson CL, Fried I, Mathern GW. Hippocampal and entorhinalcortex high-frequency oscillations (100–500 Hz) in human epileptic brain andin kainic acid–treated rats with chronic seizures. Epilepsia 1999;40:127–37.

Cascino GD. Surgical treatment for epilepsy. Epilepsy Res 2004;60:179–86.Claus S, Leijten F, Kallansee P, Klepper J, Lopes da Silva FH, Ronner H. An electro-

encephalogram beta gap after induction with diazepam: a localization methodin epileptogenic lesions. Clin Neurophysiol 2009;120:1235–44.

c high frequency oscillations. Clin Neurophysiol (2013), http://dx.doi.org/

8 S. Rampp et al. / Clinical Neurophysiology xxx (2013) xxx–xxx

Dytham C. Choosing and using statistics. 3rd ed. Hoboken, NJ: Wiley-Blackwell;.

Ebrahim ZY, DeBoer GE, Luders H, Hahn JF, Lesser RP. Effect of etomidate on theelectroencephalogram of patients with epilepsy. Anesth Analg 1986;65:1004–6.

Engel Jr J, Bragin A, Staba R, Mody I. High-frequency oscillations: What is normaland what is not? Epilepsia 2009;50:598–604.

Gancher S, Laxer KD, Krieger W. Activation of epileptogenic activity by etomidate.Anesthesiology 1984;61:616–8.

Ghoneim MM, Yamada T. Etomidate: a clinical and electroencephalographiccomparison with thiopental. Anesth Analg 1977;56:479–85.

Gigli GL, Gotman J. Effects of seizures and carbamazepine on interictal spiking inamygdala kindled cats. Epilepsy Res 1991;8:204–12.

Glass GV, Peckham PD, Sanders JR. Consequences of failure to meet assumptionsunderlying fixed effects analyses of variance and covariance. Rev Educ Res1972;42:237–88.

Goldenholz DM, Ahlfors SP, Hämäläinen MS, Sharon D, Ishitobi M, Vaina LM, et al.Mapping the signal-to-noise-ratios of cortical sources inmagnetoencephalography and electroencephalography. Hum Brain Mapp2009;30:1077–86.

Gotman J, Gloor P, Quesney LF, Olivier A. Correlations between EEG changes inducedby diazepam and the localization of epileptic spikes and seizures.Electroencephalogr Clin Neurophysiol 1982;54:614–21.

Gotman J, Marciani MG. Electroencephalographic spiking activity, drug levels, andseizure occurrence in epileptic patients. Ann Neurol 1985;17:597–603.

Harwell MR, Rubinstein EN, Hayes WS, Olds CC. Summarizing Monte Carlo results inmethodological research: the one- and two-factor fixed effects ANOVA cases. JEduc Stat 1992;17:315–39.

Hsieh JC, Shih YS, Hwang LD, Lui PK, Lui PW, Lee TY, et al. Activation of epileptogenicactivities by etomidate in electrocorticoencephalography (ECoG) duringoperation for epilepsy. Ma Zui Xue Za Zhi 1990;28:127–35.

Jacobs J, LeVan P, Chander R, Hall J, Dubeau F, Gotman J. Interictal high-frequencyoscillations (80–500 Hz) are an indicator of seizure onset areas independent ofspikes in the human epileptic brain. Epilepsia 2008;49:1893–907.

Jacobs J, LeVan P, Châtillon C, Olivier A, Dubeau F, Gotman J. High frequencyoscillations in intracranial EEGs mark epileptogenicity rather than lesion type.Brain 2009;132:1022–37.

Jawad S, Oxley J, Wilson J, Richens A. A pharmacodynamic evaluation of midazolamas an antiepileptic compound. J Neurol Neurosurg Psychiatry 1986;49:1050–4.

Jones-Gotman M, Sziklas V, Djordjevic J, Dubeau F, Gotman J, Angle M, et al.Etomidate speech and memory test (eSAM): a new drug and improvedintracarotid procedure. Neurology 2005;65:1723–9.

Kettenmann B, Feichtinger M, Tilz C, Kaltenhäuser M, Hummel C, Stefan H.Comparison of clonidine to sleep deprivation in the potential to induce spike orsharp-wave activity. Clin Neurophysiol 2005;116:905–12.

Lix LM, Keselman JC, Keselman HJ. Consequences of assumption violationsrevisited: a quantitative review of alternatives to the one-way analysis ofvariance ‘‘F’’ test. Rev Educ Res 1996;66:579–619.

Mason KP, O’Mahony E, Zurakowski D, Libenson MH. Effects of dexmedetomidinesedation on the EEG in children. Paediatr Anaesth 2009;19:1175–83.

Meinck HM, Möhlenhof O, Kettler D. Neurophysiological effects of etomidate, a newshort-acting hypnotic. Electroencephalogr Clin Neurophysiol 1980;50:515–22.

Please cite this article in press as: Rampp S et al. Etomidate activates epilepti10.1016/j.clinph.2013.07.006

Pastor J, Domínguez-Gadea L, Sola RG, Hernando V, Meilán ML, de Dios E, et al. Firsttrue initial ictal SPECT in partial epilepsy verified by electroencephalography.Neuropsychiatr Dis Treat 2008;4:305–9.

Pastor J, Wix R, Meilán ML, Martínez-Chacón JL, de Dios E, Domínguez-Gadea L,et al. Etomidate accurately localizes the epileptic area in patients with temporallobe epilepsy. Epilepsia 2010;51:602–9.

Placidi F, Tombini M, Romigi A, Bianchi L, Izzi F, Sperli F, et al. Topiramate: effect onEEG interictal abnormalities and background activity in patients affected byfocal epilepsy. Epilepsy Res 2004;58:43–52.

Rampp S, Kaltenhäuser M, Weigel D, Buchfelder M, Ingmar Blümcke I, Dörfler A,et al. MEG correlates of epileptic high gamma oscillations in invasive EEG.Epilepsia 2010;51:1638–42.

Rosenow F, Lüders H. Presurgical evaluation of epilepsy. Brain 2001;124:1683–700.Scheirer CJ, Ray WS, Hare N. The analysis of ranked data derived from completely

randomized factorial designs. Biometrics 1976;32:429–34.Siggins GR, Henriksen SJ, Chavkin C, Gruol D. Opioid peptides and epileptogenesis in

the limbic system: cellular mechanisms. Adv Neurol 1986;44:501–12.Staba RJ, Wilson CL, Bragin A, Fried I, Engel J. Quantitative analysis of high-

frequency oscillations (80–500 Hz) recorded in human epileptic hippocampusand entorhinal cortex. J Neurophysiol 2002;88:1743–52.

Stefan H, Heers M, Schmitt HJ, Rauch C, Kaltenhäuser M, Rampp S. Increased spikefrequency during general anesthesia with etomidate formagnetoencephalography in patients with focal epilepsies. Clin Neurophysiol2010;121:1220–6.

Stefan H, Lopes da Silva FH. Epileptic neuronal networks: methods of identificationand clinical relevance. Front Neurol 2013;4:8.

Vanlersberghe C, Camu F. Etomidate and other non-barbiturates. Handb ExpPharmacol 2008:267–82.

Wass CT, Grady RE, Fessler AJ, Cascino GD, Lozada L, Bechtle PS, et al. The effects ofremifentanil on epileptiform discharges during intraoperativeelectrocorticography in patients undergoing epilepsy surgery. Epilepsia2001;42:1340–4.

Worrell GA, Parish L, Cranstoun SD, Jonas R, Baltuch G, Litt B. High-frequencyoscillations and seizure generation in neocortical epilepsy. Brain2004;127:1496–506.

Wu JY, Sankar R, Lerner JT, Matsumoto JH, Vinters HV, Mathern GW. Removinginterictal fast ripples on electrocorticography linked with seizure freedom inchildren. Neurology 2010;75:1686–94.

Wyler AR, Richey ET, Atkinson RA, Hermann BP. Methohexital activation ofepileptogenic foci during acute electrocorticography. Epilepsia 1987;28:490–4.

Xiang J, Liu Y, Wang Y, Kirtman EG, Kotecha R, Chen Y, et al. Frequency and spatialcharacteristics of high-frequency neuromagnetic signals in childhood epilepsy.Epileptic Disord 2009a;11:113–25.

Xiang J, Liu Y, Wang Y, Kotecha R, Kirtman EG, Chen Y, et al. Neuromagneticcorrelates of developmental changes in endogenous high-frequency brainoscillations in children: a wavelet-based beamformer study. Brain Res2009b;1274:28–39.

Zijlmans M, Huiskamp GM, Cremer OL, Ferrier CH, van Huffelen AC, Leijten FSS.Epileptic high-frequency oscillations in intraoperative electrocorticography: theeffect of propofol. Epilepsia 2012;53:1799–809.

Zijlmans M, Jacobs J, Zelmann R, Dubeau F, Gotman J. High-frequency oscillationsmirror disease activity in patients with epilepsy. Neurology 2009;72:979–86.

c high frequency oscillations. Clin Neurophysiol (2013), http://dx.doi.org/