Volume18 Number 1 Spring 1383 May 2004
USEFULNESS OF SLEEP-DEPRIVED EEG IN THE DIAGNOSIS OF SEIZURE DISORDERS IN CHILDREN
SEYYED MOHAMMAD RAFIEI, M.D.
From the Department of Pediatric Neurology, Shiraz University of Medical Sciences, Shiraz, Iran.
The provocative effects of sleep deprivation (SD) on patients with seizure disorder and on the electroencephalogram (EEG) are well known.The purpose of the study was to test its routine use and usefulness in the pediatric and adolescent age group with, or suspect of having, seizure disorder, especially those presenting with first unprovoked seizure.
Between September 2000 and November 2002, among patients referring to the author’s clinic and the Pediatric Emergency Department of Nemazi Hospital, Shiraz, 598 children and adolescents were randomly assigned to a 10-minute surface EEG, either routinely, or after a period of partial or total SD. Those who refused or could not tolerate the SD procedure entered the routine group. The EEG’s were analyzed for the presence of epileptic abnormality.
Of 598 patients, 544( 91%) had clinical seizures; the rest (54, 9%) were seizure suspects or had seizure “mimickers” (syncope, night terrors, pseudo-seizures, etc). Of seizure patients, 210 ( 38.6 % ) suffered from a first unprovoked seizure, and the rest (334, 61.4%) had more than one seizure episode at the time of EEG performance.
Abnormal epileptiform EEG’s were more frequently seen in the group of seizure patients who were sleep-deprived before performing the EEG, as compared to the routine group. Conversely, more normal EEG’s were seen upon routine performance of EEG ( p value <00001, χ2 = 76.5).
Similarly, more abnormal sleep-deprived EEG’s were seen in patients with first unprovoked seizure than when EEG was done routinely in this subgroup of patients. Patients with complex partial seizure had significantly more abnormal EEG’s when sub- jected to SD (p<00001).
Older patients could tolerate SD better than the younger age group; so more routine EEG’s had to be performed in the younger age group.Younger children could only tolerate partial SD. There was no difference in the yield of EEG in patients with absence epilepsy between the two groups. In 5 patients with juvenile myoclonic epi- lepsy, SD precipitated a generalized tonic-clonic seizure which necessitated acute anti- epileptic drug administration to the seizure activity.
All patients with non-seizure episodes showed normal EEG’s.
In conclusion SD can be safely applied in the work-up of selected pediatric patients with seizure disorder, especially in those with their first unprovoked seizure, in whom an abnormal EEG might dictate anti-epileptic drug therapy.
Not all children with seizure disorder need to undergo SD for EEG, as it may not 21
Sleep-Deprived EEG for Seizure Disorder Diagnosis in Children
be necessary in some, and intolerable in others. In non-seizure episodes, if used in appropriate pediatric age groups, it can be very helpful to ensure that a seizure disorder is not present.
MJIRI, Vol. 18, No. 1, 21-27, 2004.
Keywords: Seizure Disorder, Epilepsy, Electroencephalography(EEG), Sleep Deprivation (SD).
EEG is the most useful procedure in patients with suspicion of seizure disorder. Obtaining an EEG after a period of sleep deprivation (SD) is a well known proce- dure to increase the yield of EEG “positivity” and bring about the inter-ictal epileptiform activity.1-10 It is some- times used as an activation procedure in a patient with seizure disorder whose first EEG has been negative; it is also recommended to be used routinely in children and adults11,26 for the diagnosis of seizure disorder, or to in- duce sleep in children for obtaining sleep EEG’s.11 Espe- cially in a child with first unprovoked seizure, a positive EEG helps in determination of the risk of seizure recur- rence, and the decision of drug therapy.12 In children
Table I. Patient characteristics.
with “seizure mimickers”–syncope,breath holding spells, emotional problems, pseudo-seizures–13 a negative EEG would be reassuring both to parents and physician, and obviates unnecessary anti-epileptic therapy. SD may activate interictal epileptiform discharges and provoke clinical seizures in susceptible individuals14 and is in- convenient or not feasible in infants and small children.15
The aim of the study was to see which patients ben- efit most from obtaining an EEG with SD, and whether it should be part of a routine EEG in every infant or child with or suspect of seizure disorder in our region.
MATERIAL AND METHODS
From September 2000 to November 2002, 598 infants,
Total no. 598 Male 356 M/F 1.47 :1 Female 242
NL Exam 597 Age (in year):
<5 >5 <10 >10 <15 >15
Abnormal. Exam 1 # Age
# 38 88 26
34 34 25
157 175 46
Type of Seizure GTCS 364 Gen. atonic GTCS/jerk (J.M.E.) CPS 30 TOTAL
Faints Night Terrors Anxiety Reaction Depression
Gen. tonic 26 Rol +/- 2 Gen.
31 Absence +/- GTCS
21 Pseudo-Seizure 7 Vertigo
2 R/O Absence 2 Sleep Jerks
Table II. EEG abnormality vs. patient’s condition at EEG performance.
Table III. Frequency of EEG abnormality vs. number of seizures before EEG performance.
Condition at EEG No. Normal EEG Abnormal EEG
Routine Awake Asleep
Awake Asleep Awake—>Asleep Total
253 154 99 27 18 9 0 0 0 280 172 108
133 50 83 61 8 53 70 6 64 264 64 200
First Seizure N= 210 Normal EEG Abnormal EEG
>1 Seizures N = 334 Normal EEG Abnormal EEG
Routine Awake 91 Asleep 6 Awake —>Sleep 0
Total 97 Sleep-Deprived
64 27 5 1 0 0 69 28
Awake 46 Asleep 42 Awake —>Sleep 25 25
20 26 6 36
162 90 72 21 13 8 000 183 103 80
87 30 57 19 2 17 45 6 39
151 38 113
children, and adolescents, were randomly selected for the study from the author’s private clinic patients, as well as those brought in to Nemazi Hospital Pediatric Emergency Room, Shiraz, with a chief complaint of one or more clinical seizure episode(s), or paroxysmal be- haviors suspect of seizure disorder, for ruling out a sei- zure disorder. A thorough clinical evaluation, including detailed history and general physical and neurological examination was performed on each patient. Except for one case, patients with abnormal neurological examina- tion were excluded from the study.
Patients were randomly advised to undergo either 1) a routine EEG, which according to the patient’s age and cooperation, included: a wake, sleep, or awake -to-sleep tracing with or without a 3-minute hyperventilation and 1-30 Hz intermittent photic stimulation, or, 2) the same EEG after a period of SD.
The second group were advised to not go to sleep either for 24 hours before performing EEG (total SD),9 or in the case of younger children, to be kept awake as much as possible in the night before EEG performance, and to sleep only for no longer than 3 hours, and then be awakened (partial SD), and be present for the EEG laboratory, to undergo the EEG procedure, exactly as mentioned above.
The EEG machine was a 10-channel Nihon-Codon (Ja- pan) paper EEG machine, as was shown in our previous report16 to be effective in showing epileptiform activity. Sleep was achieved either normally, or if needed, after a hypnotic, mainly promethazine syrup 0.5 mg/kg. In those who could not tolerate, or refused, the SD, the EEG was obtained routinely at a later time.
A 10-minute EEG was thus obtained. It was visually ana- lyzed for the presence of: 1) epileptic abnormality, i.e., spikes
Sleep-Deprived EEG for Seizure Disorder Diagnosis in Children
Table IV. Type of seizure vs EEG abnormality.
GTCS : Generalized tonic-clonic seizures ROl.: Rolandic seizures J.M.E. : Juvenile myoclonic epilespy
Table V. Seizure “mimickers’ and EEG at routine vs sleep deprivation.
R1: Routine, awake; R2: Routine, asleep; R3: Routine, awake ——> sleep SD1: Sleep-Deprived, Awake; SD2: SD, Asleep; SD3: SD, awake —> asleep
Seizure Type Routine EEG Sleep Deprived EEG Normal Abnormal Normal Abnormal
Absence +/- GTCS Gen.Atonic CPS
GTCS J.M.E. Rol +/- GTCS
2 13 10 6 4 6 10 83 19 139 67 42 116 46 615 38 320 107 710
All EEG’s taken were normal
Condition No. EEG: R1 R2 R3 SD1 SD2 SD3
Faint R/O Absence Anxiety Pseudo-Seizures Night Terrors Depression SleepJerk Vertigo
21 7 10 4
2 2 12 5 7 2 2 6 3
2 2 5
7 2 2
& waves, poly spikes and waves, sharp-&-slow waves, 2) non-epileptic abnormality, i.e., generalized or focal slowing.17 A data analysis was done.
Table I shows patient’s characteristics. Of 598 in-
fants, children, and adolescents, 544(91%) had some form of clinical seizure activity, for the first time (210, 38.6%), or more than one seizure episode (334, 61.4 %), before being subjected to EEG procedure. 54 patients (9%) had non-seizure paroxysms, most commonly syncopal at- tacks. Except for one patient, all other patients had nor- mal developmental and neurological evaluation. This
patient, a 14 year old boy with congenital hemiparesis and mild developmental delay with two episodes of noc- turnal generalized tonic-clonic seizures(GTCS), under- went a sleep-deprived EEG, first awake and then went to sleep during the EEG procedure, and his EEG proved abnormal. Of all patients with seizure disorder, GTCS pre- dominated (364,66.9%) (Tables I, IV).Other seizure types had equal frequencies.
Routine EEG’s predominated somewhat over sleep- deprived ones; this was because some patients, espe- cially those below 5 years, either refused, or could not tolerate sleep deprivation; so they underwent a routine EEG procedure.
As shown in Table II, the number of normal EEG’s predominated in the routine group (172 vs. 64); con- versely, in the sleep-deprived group, more abnormal EEG’s were encountered (200 vs. 108) (p value <0.0001)
In the routine group, more EEG’s had to be performed in awake state (253 vs. 133); in the sleep-deprived group,more sleep,and awake-to-sleep EEG’s could be performed (131 vs. 27).
As shown in Table III, when patients with first sei- zure underwent a sleep-deprived EEG, more abnormal results were obtained (87 vs. 28); the same prevailed for patients with more than one seizure at the time of EEG performance (113 vs. 80). Slightly more sleep-deprived EEG’s were performed on patients with first seizure (113 vs. 97); this may also partly account for more abnormal EEG’s in this group.
When patients were sleep-deprived, more awake-to- sleep EEG’s could be performed (70 vs. zero); the major- ity of these EEG’s were abnormal ( 64 out of 70) ( Table II). This procedure combines the “power” of sleep deprivation and sleep in provoking epileptiform dis- charges.21
When first-seizure group was compared to multi-sei- zure group regarding EEG abnormality, the second group had more abnormal EEG’s whether or not sleep-deprived (80 vs. 28; 113 vs. 87).
When seizure type is tested versus frequency of EEG abnormality relative to condition at EEG performance, it seems that, a) for absence seizure it does not matter whether it is performed routinely or with SD; both show high percentage of EEG abnormality ( Table IV); b) in patients with generalized seizures, SD did increase the yield of EEG, whether or not primarily or secondarily gen- eralized (p< 0.0001). This is as well true for complex par- tial seizure (CPS), which showed high yield with SD (Table IV) (p< 0.0001); c) regarding juvenile myoclonic epilepsy (J.M.E.), although SD increased the yield, it provoked clinical GTCS in 5 patients, which necessitated acute anti-epileptic therapy, in the form of diazepam rectally or intravenously.
Regarding the age at which sleep SD could best be
implemented, the older the child the easier it could be done, so most of the patients sleep-deprived were more than 5 years old. In younger children, only partial sleep deprivation could be performed.
SD – even partial - facilitated putting patients to sleep, and obviated the need to sedate the patients artificially. Hyperventilation (H.V.) increased the EEG abnormal- ity only in patients with absence seizure, equally in both groups; other patients in either group did not show any increase in EEG abnormality. This may reflect ineffective
H.V. by the uncooperative or sleepy patients. Early in the study, equal numbers of patients (n=4) in either group showed increase in EEG abnormality with photic stimulation (P.S.). Later, P.S.did not enhance any further EEG abnormality in either group. So the differen- tial effect of P.S. in the two groups was not further stud-
ied. All patients with “seizure mimickers” had normal
EEG’s when sleep-deprived (Table V).
Obtaining EEG’s using sleep SD is not a recent pro- cedure. Rodin et al. noted high voltage paroxysmal ac- tivity in the EEG’s of 16 normal subjects following 120 hours of SD.1 Cases of convulsion in normal subjects (pilots and soldiers) after SD were later reported.2-5 SD has a very low false-negative rate (1.2-2.2 %).2,4,6,7 SD has activating effects on interictal epileptiform dis- charges over and above the effects of sleep per se.8,10,27,28 It may obviate the need for a second routine EEG when the first one is normal.23,24,25
SD may also provide a useful way of inducing sleep in children, avoiding sedative artifacts on the EEG back- ground.11 However, the effect of SD on EEG is indepen- dent of the effect of sleep itself.10
As is well known, sleep unravels or potentiates epi- leptiform activity on EEG.14 Activating effects of SD may be more pronounced in children compared to adults.20,22
So it seems prudent to obtain a sleep deprived awake EEG followed by sleep, in children with clinical impres- sion of seizure activity, or paroxysmal behavior for which seizure disorder is to be ruled out. This obviates the need to perform a second EEG in those whose routine EEG has been normal, and the suspicion of clinical sei- zure is still high. Potential “side effects” of SD include:15 stress and inconvenience of SD, especially in children, potential accident to patient when en route to EEG labo- ratory due to sleepiness; this means that another per- son should accompany the patient, which implies loss of work or school for two persons, and minor chance of seizure provocation.
Patients with known seizure disorder experience in- crease in seizure frequency when sleep-deprived. Acti-
Sleep-Deprived EEG for Seizure Disorder Diagnosis in Children
vation of seizure discharges, and increase in seizure af- ter SD is a known phenomenon.14
Patients with juvenile myoclonic seizure may be par- ticularly susceptible to the seizure-provoking effect of SD, and in them, GTCS activity may be precipitated if sleep-deprived,30 as shown in this study. Therefore, a great precaution is advised when counseling these pa- tients regarding obtaining an EEG with SD.
The EEG laboratory and parents should be provided with necessary information and guidance to abort the seizure, if one occurs in the sleep deprived patient be- fore, during or after the EEG procedure. It is probably better to admit these patients at hospital, sleep-deprive them in the ward, and perform the EEG with close moni- toring. Or, if the history is highly suggestive of juvenile myoclonic epilepsy, to perform a routine EEG, or start anti-epileptic drugs without performing EEG.
So based on our results, we recommend a sleep de- prived EEG in children and adolescents who present with first unprovoked seizure.
Sleep deprivation may be inconvenient for children,15 and may precipitate clinical seizure in patients with sei- zure disorder.14,30 GTCS activity may be precipitated if the patient is sleep - deprived.30 Children with absence seizures might not need a sleep–deprived EEG, as the routine awake EEG with hyperventilation usually unrav- els the typical EEG abnormality in this group of patients. Patients with juvenile myoclonic epilepsy may be ex- empted from SD before performing EEG, as it may pro- voke a clinical GTCS.
In infants and children and adolescents who are “sei- zure mimickers”,13 a negative sleep deprived EEG may be more assuring to the parents and clinicians, and obvi- ates unnecessary and potentially harmful anti-epileptic administration.
Infants and children below 5 years of age may be partially sleep-deprived, merely to induce natural sleep and obviate the need for sedation, which may impart artifact on EEG tracing.
With minor exceptions, SD can safely be employed in the protocol for EEG recording in children. It is espe- cially useful in children with first seizure episode, when an abnormal EEG result may dictate initiation of anti- epileptic therapy.
Patients with absence epilepsy may not need it; in pa- tients with juvenile myoclinic epilepsy, it is advisable not to employ it for the potential danger of provoking a clinical GTCS episode. In younger children partial sleep SD is advised.
The author is especially thankful to the staff of
Motahhari Clinic EEG Laboratory for their preparation of the patients’ EEG’s and to Dr Sarafraz and Ms Baghaee for their guidance in statistical analysis.
1. Rodin E, Luby ED, Gottlieb JS: The electro-encephalogram during prolonged experimental sleep deprivation. EEG & Clinical Neurophysiology 14: 544-551, 1962
2. Bennet DR: Sleep deprivation and major motor convul- sions. Neurology 13: 953-958, 1963.
3. Bennet DR, Mattson RH, Zitter FA et al: Sleep depriva- tion: Neurological and electroencephalographic effects. Aerospace Medicine 35: 888-890, 1964.
4. Bennet DR, Zitter FA, Lisk EA: Electroencephalographic effects of sleep deprivation in flying personnel. Neurology 19: 375-379, 1969.
5. Gundeson CH, Dunne PB: Sleep deprivation seizures. Neu- rology 23: 678-686, 1973.
6. Geller MR,Gourdji N, Chisthoff N, et al :The effects of sleep deprivation on the EEGs of epileptic children. Dev Med Child Neurol 11: 771-776, 1969.
7. Welch LK, Stevens JB: Clinical value of the electroencepha- logram following sleep deprivation. Aerospace Medicine 42: 349-351, 1971.
8. Fountain NB, Kim JS, Lee SI: Sleep deprivation activates epileptiform discharges independent of activating effects of sleep. J Clin Neurophysiology 15: 60-75, 1998.
9. Neidermyer E: Sleep and EEG. In: Neidermyer E, Da Silva FL, (eds). Electroencephalography: Basic Principles, Clini- cal Applications, and Related Fields. Fourth Edition, Phila- delphia: Williams & Wilkins Company, pp. 175-176, 1999.
10. Ellingson RJ, Wilken K, Bennett DR: Efficacy of sleep deprivation as an activating procedure in epilepsy patients. J Clin Neurophysiol 15: 69-75, 1984.
11. Liamsuwan S, Gratten-Smith G, Fagan E, Bleasel A, Anthony J: The value of partial sleep deprivation as a routine measure in pediatric electroencephalography. Jour- nal of Child Neurology 15(1): 26-29, 2000.
12. Hirtz D, Ashwal S, Berg A, et al: Practice parameter: evalu- ating a first nonfebrile seizure in children. Report of the Quality Standards Subcommittee of the American Acad- emy of Neurology, the Child Neurology Society, and the American Epilepsy Society. Neurology 55(September): 616-623, 2000.
13. Haslam RH: Conditions that Mimic Seizures. In: Behrman RE, Kliegman RM, Jenson HB, (eds). Nelson Textbook of Pediatrics, 16th Edition, Philadelphia:W.B. Saunders Com- pany, pp. 1829-1832, 2000.
14. Kotagal P: The relation between sleep and epilepsy. Semi- nars in Pediatric Neurology 8(4) (December): 241-250, 2001.
15. Glick TH: The sleep-deprived electroencephalogram: evi- dence and practice. Archives of Neurology 59 (8): 1235-