Epilepsy Research 70 (2006) 83–88

Short communication

Seizure anticipation by patients with focal and generalizedepilepsy: A multicentre assessment of premonitory symptoms

Andreas Schulze-Bonhage a,∗, Christoph Kurth b, Astrid Carius a,Bernhard J. Steinhoff b, Thomas Mayer c

a Epilepsy Centre, University Hospital Freiburg, Germanyb Epilepsy Centre, Kork, Germany

c Epilepsy Centre, Kleinwachau, Germany

Received 24 October 2005; received in revised form 22 January 2006; accepted 6 February 2006Available online 10 March 2006

Abstract

Purpose: To assess subjective seizure anticipation in patients with focal and generalized epilepsy.Methods: Five hundred consecutively recruited out-patients (251 male, 249 female, mean age 38.1 year) from three Germantertiary epilepsy referral centres filled out questionnaires regarding subjective anticipation of seizures by at least 30 min and totiming and semiologic characteristics of their premonitory symptoms versus those of ictal phenomena. Patients were not regardedas having prodromi if the semiology of symptoms reported long before a seizure was identical to auras.Results: 6.2% of patients reported that they were able to anticipate seizures. Premonitory symptoms were classified as stereotypedin all but one patient. An intraindividual semiologic analysis showed that the majority of these patients had symptoms, whichwere distinct from ictal experiences during auras. Seizure anticipation was reported both by patients with focal and idiopathicgeneralized epilepsy. The median estimated time interval between occurrence of premonitory symptoms and seizure onset was90 min.Conclusions: This study gives evidence that both patients with focal and idiopathic generalized epilepsy may subjectivelyanticipate the occurrence of epileptic seizures. Premonitory symptoms are distinct from auras in terms of semiology and timeof occurrence. The lower percentage of patients regarded as having premonitory symptoms as compared to some earlier reportsmay be related to stricter criteria and to the exclusion of auras, which could directly evolve into seizures, and other ictal events.Premonitory symptoms occur at similar periods prior to seizures as anticipatory EEG-changes have been reported using methodsfrom time series analysis.© 2006 Elsevier B.V. All rights reserved.

Keywords: Premonitory symptoms; Prodromi; Focal epilepsy; Generalized epilepsy; Seizure anticipation

∗ Correspondence to: University Hospital, Freiburg, Breisacher Street 64, D-79106 Freiburg, Germany. Tel.: +49 761 270 5425;fax: +49 761 270 5003.

E-mail address: schulzeb@nz.ukl.uni-freiburg.de (A. Schulze-Bonhage).

0920-1211/$ – see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.eplepsyres.2006.02.001

84 A. Schulze-Bonhage et al. / Epilepsy Research 70 (2006) 83–88

1. Introduction

The ability of some patients to anticipate the occur-rence of epileptic seizures has fascinated physiciansfrom the time of Hippocrates. In his famous work“On the sacred disease” he considered the occur-rence of subjective phenomena preceding seizuresas quite common: “But such persons as are habitu-ated to the disease know beforehand when they areabout to be seized and flee from men . . . little chil-dren at first fall down wherever they may happento be, from inexperience. But when they have beenoften seized, and feel its approach beforehand, theyflee to their mothers, or to any other person theyare acquainted with, from terror and dread of theaffection . . . ”(1). In 1885, Gowers subclassified suchwarning symptoms into epileptic “auras” and “pro-dromi” (Gowers, 1885): whereas auras represent ongo-ing epileptic activity restricted to parts of the brainwhich do cause subjective symptoms only, prodromiare poorly understood in their nature and may precedea seizure for a longer period of time. Thus, auras arelimited to patients with focal epilepsy, whereas thereis little information on this available with regard toprodromi.

Epileptic auras have been the subject of intensivesairIpla

q

(

(

(

(

2. Methods

Five hundred subsequent adult patients attendingthe epilepsy out-patient clinics were prospectivelyincluded in this study at the epilepsy centers at Freiburg(2 0 0), Kork (2 0 0) and Kleinwachau (1 0 0). Patientsincluded had to be able to fill out a questionnaire regard-ing the presence of any type of premonitory symptomson which they could base seizure anticipation, andto give their informed consent to participate in thisstudy as part of the recruitment of a study approvedby the local ethic’s committee. Two hundred and fifty-one patients were male, 249 were female, the meanage was 38.1 ± 14.7 years. Patients were classified ashaving focal or primarily generalized epilepsy basedon history, seizure semiology, EEG and imaging find-ings. In the sample, 415 had focal epilepsy, 82 hadprimarily generalized epilepsy, and 1 patient had evi-dence of both, focal and generalized epilepsy. Twopatients could not be classified based on the availabledata.

The definition of prodromi/premonitory symptomswas based on (1) the time interval of a symptom pre-ceding seizures and (2) on the fact that their semiol-ogy could be distinguished from the habitual seizuresemiology. In a questionnaire, patients were asked iftasoisPpowawwt

dar

at

tudies establishing electroclinical relationships andnalysis of the localizing value of auras particularlyn the context of presurgical monitoring in patientsegarded as possible candidates for epilepsy surgery.n contrast, there are only few systematic studies in therevalence of prodromi, their relationship to the under-ying epilepsy syndrome, and their relationship withuras.

The study reported here thus addresses the followinguestions:

1) Do patients with both focal and primarily gener-alized epilepsy report that they can anticipate theoccurrence of seizures?

2) Does the semiology of premonitory symptomsclearly differ from subjective experiences duringictal events?

3) What is the prevalence of prodromi in focal andprimarily generalized epilepsy?

4) What is the time interval patients report for seizureanticipation?

hey could anticipate seizures for at least 30 min, inccordance with a study by Hughes et al. (1993). Ifo, they were asked to give a free report on the typef symptoms, to state if they are stereotyped or not,f they occur always or only at times, and how longuch symptoms would on average precede a seizure.atients stating they would have symptoms which mayrecede a seizure by more than 30 min but normallyccur considerably shorter than 30 min prior to seizuresere not regarded to have prodromi. In addition,semiologic comparison of premonitory symptomsith symptoms experienced during an epileptic seizureas performed (see Appendix A: contents of ques-

ionaire).These data were correlated with the epilepsy syn-

rome, with the presence or absence of epileptic aurasnd other habitual seizures, and with the symptomseported during such ictal events.

The terms “prodromi”, “prodromal symptoms”nd “premonitory symptoms” are used synonymouslyhroughout the article.

A. Schulze-Bonhage et al. / Epilepsy Research 70 (2006) 83–88 85

Fig. 1. Number of patients with a given estimated time interval ofpremonitory symptoms preceding a subsequent seizure. Three addi-tional patients indicated “several hours” before a seizure, which doesnot allow for representation in this graph.

3. Results

35/500 patients (7.0%) reported that they couldanticipate seizures by more than 30 min; 30 of thesehad symptomatic focal epilepsies (prevalence 7.2%),and five (prevalence 6.1%) had idiopathic generalizedepilepsies. Twenty-five patients were able to give dataon the time premonitory symptoms preceded a seizure.The median estimated time of occurrence of prodomiprior to a seizure was 90 min (range 30 min–24 h). 9/25estimated a time period of 30–60 min, 10/25 a period of1–3 h, and 6/25 a time period of more than 3 h (6–24 h,see Fig. 1).

The most common types of prodromal symptomsare listed in Table 1; in addition, diverse symptoms likesleeplessness, psychic symptoms like feeling of strain,

Table 1Type of premonitory symptoms reported by several patients, andrespective number of reports

Premonitory symptom Number of patientsreporting the symptom

Restlessness 10Headache 6Malaise 5Nausea 2IDT

tension, irritability, nervousness, and changes in per-ception were reported in single patients. All patientsexcept one with idiopathic generalized epilepsy statedthat the semiology of prodromi was stereotyped. 17/30patients with focal epilepsies who reported premon-itory symptoms more than 30 min in advance of aseizure also had epileptic auras. In three of them, thereported semiology of auras was so similar that a cleardistinction was not possible. These cases were regardedas possibly experiencing prolonged auras or additionalauras occurring at some time interval before the onsetof an overt seizure. In the other patients with both aurasand prodromi, the reported semiology differed consid-erably (Table 2).

One out of five patients with idiopathic generalizedepilepsy with absences and generalized tonic-clonicseizures reported episodes of absentmindedness, whichpreceded generalized tonic-clonic seizures; a clear dis-tinction of these episodes from absence seizures wasnot possible in this case. If these patients with premon-itory symptoms of uncertain type are excluded from theprodromi group, 31/500 patients (6.2%) remained withsymptoms regarded as distinct prodromi (6.5% in focalepilepsy, and 4.9% in idiopathic generalized epilepsy).

The rate of occurrence of premonitory symptomswas classified as either “always” or “frequently” pre-cpsqmpotpt

4

ntawtsh

mpaired concentration 2izziness 2iredness 2

eding a seizure by the patients. 7/31 patients reportedremonitory symptoms to occur always preceding aeizure, 24/31 patients reported them to occur fre-uently but not always thus indicating that in theajority of patients, not all seizures were preceded by

remonitory symptoms. One patient with “frequently”ccurring premonitory estimated a percentage of 50%,hree others mentioned “sometimes” indicating that theercentage of seizures preceded by premonitory symp-oms was below 50%.

. Discussion

Considering the very limited knowledge about theature of premonitory symptoms, criteria for symp-oms to be classified as prodromi remain to some degreerbitrary. Here we used a rather conservative definitionith regard to the time period by which symptoms had

o precede seizures, and with regard to the subsequentemiological analysis. Certainly some patients may notave been included which had symptoms preceding

86 A. Schulze-Bonhage et al. / Epilepsy Research 70 (2006) 83–88

Table 2Premonitory symptoms and auras in those 15 out of 17 patients with focal epilepsy who were able to report details of both subjective experiences

Premonitory symptom ≥30 min before onset of clinicalseizure signs

Type of aura

General malaise, bad temper Epigastric aura, then acoustic aura with distorted hearingGeneral malaise, restlessness Psychic aura (fear,deja-vu), gustatory aura (strange flavor)Smell, restlessness, hearing voices like from a tape, feeling

people were far away when speakingAcoustic aura of similar semiology

Limb aching, tiredness, headache MalaiseHyperactivity, mentally busy, later “light breeze” passing

through the bodySensible aura in the left half of the throat like frosted by a cooler spray,evolving into a psychic aura with fear, tachycardia

Restlessness, complex hallucinations Epigastric aura“Strange feeling” Epigastric auraMild dizziness DizzinessTiredness, restlessness, impaired concentration Olfactory aura, then psychic aura (panic) accompanied by tachycardiaStrong headaches,“bad thoughts”, unpleasant feelings Headache, psychic aura (panic, stereotyped thoughts)Pressure on the temples, restless bowel, mild dizziness Cephalic auraNausea Epigastric auraRestlessness Sensible aura (often simple partial sensory-motor seizures)Restlessness Epigastric auraUnspecified changes in perception Epigastric aura, psychic aura (fear), dizziness

Premonitory symptoms which were regarded as closely resembling aura-symptoms and were thus discarded for the quantitative analysis aregiven in italics.

seizures longer than a typical aura would last; how-ever, all patients excluded from the prodromal groupbecause of symptoms that would most often occur lessthan 30 min prior to a seizure reported intervals ofseconds to minutes. This strongly suggests that thesesymptoms represent auras, which also occurred in iso-lation at longer time periods before other seizure types.

Based on these criteria, the study displays newaspects of subjective seizure anticipation with regardto their relationship to epilepsy syndromes, semiolog-ical differences with auras, and with regard to theirprevalence.

The prevalence of reported prodromi in our patientsample was lower than what might have been expectedfrom two previously published studies. In a sample of148 patients, 86% of which had focal epilepsy, Hugheset al. (1993) reported premonitory symptoms to occurwith a prevalence of 29% versus ca. 6–7% in our study;all their patients reporting prodromi had focal epilepsy.Like in our study, premonitory symptoms were definedas symptoms preceding seizures by at least 30 min, andemotional changes were most common with irritabil-ity and depression being most frequent. There is noevidence of a strikingly different patient selection inboth studies. However, differences in data acquisitionmay have contributed to the lower prevalence of pre-

monitory symptoms reported in our study. Whereas ourstudy used a questionnaire filled out by the patientalone, the latter study used an interview guided bythe physician. The spontaneous report by the patientmay not include premonitory symptoms, which thephysician but not the patient has in mind; e.g., noneof our patients reported depression as a preictal phe-nomenon he was aware of, whereas this was the thirdmost common symptom in Hughes’ study, and has beenreported in other studies as a sign which may precedeseizures for longer periods like 24 h (e.g., Blanchet andFrommer, 1986). From the method’s description of thatstudy it is not clear if the interview was standardized,and to which degree the interviewer actively checkedfor expected symptoms, which might have increasedthe rate of premonitory symptoms.

Hughes et al. found premonitory symptoms onlyin patients with focal epilepsies. This could suggestthat they have some commonality in their generationwith auras, e.g., might be related to changes in theEEG dynamics of the epileptic focus. However, boththeir investigation and our study found semiologic dif-ferences between the reported prodromi and epilep-tic seizures, and prodromi were usually not reportedto occur as a paroxysmal phenomenon with clear-cut onset and end as it is the case in epileptic auras.

A. Schulze-Bonhage et al. / Epilepsy Research 70 (2006) 83–88 87

We found premonitory symptoms not only in patientswith focal epilepsy but also with idiopathic general-ized epilepsies, even if their prevalence may be slightlylower. Taking into consideration that the study ofHughes et al. included only 20 patients with idiopathicgeneralized epilepsy in their sample, their presence ata prevalence in the range of 5% could have easily beenmissed. The presence of premonitory symptoms alsoin non-focal epilepsies is, however, of considerableimportance with regard to possible pathomechanismsunderlying their generation, and strongly supports thenotion that prodromi represent a distinct entity of symp-toms not related to focal ictal activity.

In the Hungarian multicentre study on warnings andinitial symptoms reported by Rajna et al. (1997), a ques-tionnaire was used to assess warning symptoms. In thisstudy, 262/562 patients (47%) reported to experienceany type of symptom preceding a seizure. In 57% ofthose, symptoms occurred only within a period of 5 minbefore seizure onset, and would thus not commonly beclassified as a prodrome. There are no data given on thetime interval between the occurrence of warning symp-toms and seizure onset in the remaining 20% of thetotal patient sample leaving open the question whethersuch symptoms had to be regarded as epileptic aurasor premonitory symptoms in the strict sense used inoshm

tptlwe

pcosdrfi

ii

Their classification as premonitory symptoms is ratherrelated to the temporal relationship to the occurrenceof seizures. Some of the symptoms could in principlealso occur as side effects of antiepileptic medication;it would, however, not be plausible to assume thatseizures would occur particularly frequent followingperiods of high CNS levels with signs of CNS intox-ication, and there were several patients who reportedthat they would take additional drugs in case that pre-monitory symptoms occur in order to prevent a seizure.This strongly argues against a drug overdose as ansufficient explanation of the premonitory symptomsreported.

The ability of patients to anticipate the occurrenceof seizures has recently gained considerable interestdue to the fact that this might open up new treatmentoptions. Both behavioral strategies for seizure controland classical medical interventions could be used in thesubset of patients having premonitory symptoms. Dataon prodromi are sparse, which is in striking contrast tothe fact that they have been mentioned in textbooks onepilepsies from the time of Gowers on. By differentiat-ing symptoms of auras and prodromi and by correlatingpremonitory symptoms to the epilepsy syndrome, ourstudy gives additional support for the notion that pro-dromi have a particular underlying physiology, and thatsetobbeoipiomapuobsamp

ur study. In addition, the reported types of warningymptoms included both those found in our study likeeadache, and others like epigastric sensations, whichay be presumed to at least partially represent auras.Unfortunately the report of the Hungarian multicen-

re study did not separately report on patients with idio-athic generalized epilepsy. It is only mentioned thathey occurred “primarily in focal epilepsies”, whicheaves open the question if there were also patientsith prodromi in well-defined idiopathic generalized

pilepsies.The presence of prodromi also in patients with idio-

athic generalized epilepsy has been mentioned in aongress report by Giuccioli et al. (1990) on a groupf hospitalized epilepsy patients. Although exact inclu-ion criteria for classification of a symptom as a pro-rome are not given, the prevalence was reported aselatively low (8.6%), thereby closely resembling ourndings.

The prodromal symptoms reported by the patientsn any of these studies cannot be assumed to be specificn the sense that they occur only preceding a seizure.

ubjective seizure anticipation occurs not only in focalpilepsies. Further analysis with regard to the predic-ive value (sensitivity and specificity) of prodromi aref particular interest as premonitory symptoms mighte related to changes in brain dynamics, which haveeen reported to precede seizures in patients with focalpilepsies (e.g., Mormann et al., 2003, 2005). Althoughur data do not allow to prove any such correlationship,t is of particular interest that the time of occurrence ofremonitory symptoms is in a similar range as changesn EEG dynamics have been reported in some studiesn EEG-based seizure anticipation. Patients with pre-onitory symptoms may be of particular interest for

n investigation of electroclinical correlates of thesehenomena the nature of which has so far remainednclear. Although our study suggests that the numberf patients which are able to anticipate seizures maye lower than suspected in earlier studies, the differentemiologic characteristics of prodromi as compared touras and other seizure types in the majority of patientsakes an investigation of their physiological correlates

articularly interesting.

88 A. Schulze-Bonhage et al. / Epilepsy Research 70 (2006) 83–88

Appendix A. English translation of thequestionnaire

Dear Patient,At the epilepsy centre of Freiburg, precendents of

seizures felt more than 30 min before a seizure arebeing investigated. We would be thankful if you wouldanswer the following questions according to your per-sonal experience. All data will be kept strictly confidentand not transferred to others.

1. I do feel more than half an hour before a seizure thatit is going to occur (no/yes,often/yes, always).

If you have answered with “yes”, please fill outthe following questions

2. The feeling I have in this case is always similar(yes/no).

3. As a premonitory symptom, I feel . . ..4. In general, I do feel this some . . . hours before a

seizure.5. If I feel premonitory symptoms, I do prepare myself

for a seizure to come (no/yes: how? . . .).

References

Hippocrates: on the sacred disease. Translated by Francis Adams,2006.

Blanchet, P., Frommer, G.P., 1986. Mood change preceding epilepticseizures. J. Nerv. Ment. Dis. 174, 471–476.

Giuccioli, D., Czuczwara, H., Finkler, J., Leitenberger, J., May, T.,Nothbaum, N., Wolf, P., 1990. Epilepsie und Prodromi: eineprospektive Untersuchung. In: Epilepsie 89. Einhorn-Presse Ver-lag, Reinbeck, pp. 312–321.

Gowers, W.F., 1885. Epilepsy and Other Chronic Convulsive Disor-ders. William Wood & Co, New York.

Hughes, J., Devinsky, O., Feldmann, E., Bromfield, E., 1993. Pre-monitory symptoms in epilepsy. Seizure 2, 201–203.

Mormann, F., Kreuz, T., Andrzejak, R.G., David, P., Lehnertz, K.,Elger, C.E., 2003. Epileptic seizures are preceded by a decreasein synchronization. Epilepsy Res. 53, 173–185.

Mormann, F., Kreuz, T., Rieke, C., Andrzejak, R.G., Kraskov,A., David, P., Elger, C.E., Lehnertz, K., 2005. On the pre-dictability of epileptic seizures. Clin. Neurophysiol. 116, 569–587.

Rajna, P., Clemens, B., Csibri, E., Dobos, E., Geregely, A., Gottschal,M., Gyorgy, I., Horvath, A., Horvath, F., Mezofi, L., Velkey, I.,Veres, J., Wagner, E., 1997. Hungarian multicentre epidemiologicstudy of the warning and initial symptoms (prodrome, aura) ofepileptic seizures. Seizure 6, 361–368.