Infantile spasms: Ictal phenomena

Original Articles

Infkntile Spasms: Ictal Phenomena Don k. King, MD*, Paul R. Dyken, MD*f, Ira L. Spinks,Jr*, and Alice J. Murvin, RN*

Ictal phenomena were studied during three separate sixhour videolpolygraphic recording sessions in 10 patients with infantile spasms; 1,079 spasms occurted. Frequency during wakefulness (7.7 spasms/hour) was greater than that during sleep (2.5 spasms/hour); 46.6% of spasmsoccurted in clusters. Spasmswere composed of one or more of three phases:a myoclonic contraction, a tonic contraction, and/or an arrest of activity. The most common types were myoclonic-tonic (40.3%) and myoclonic alone (36.3%). When classified by postural motor phenomena, 41.6% were “flexor”, 16.3% “extensor”, 39.0% “mixed”, and 3.1% ‘ ‘arrest’ ’ alone. Electrographic monitoring revealed that myoclonic contractions were associated with an initial paroxysmal event. Tonic contractions and arrestswere usually associatedwith suppressionof ekctroencephalographic activity with or without rhythmic activity. Knowledge of these clinical and electrographic features is important for diagnosis and evaluation of proposed treatments.

to confirm and extend these observations by analyzing the temporal occurrence, ictal phenomena, and electrographic characteristics of spasmsin ten patients with this syndrome.

Methods Patients entered

included

in the present

treatment

of refractory

responded

to previous

antiepileptic

drugs.

infantile trials

The ten patients of entry months,

mean

recognized

the

Neurology; Mobile,

‘Department

College

University AL

of Neurology

of Georgia;

Augusta,

of South

and the *Nursing

Scrvicc;

GA and the fDcparrmcnr

clonic

seizures,

ctiologic

Alabama

College

of Mcdicinc;

pyridoxinc,

and standard had undergone

age when

two

and

one

weeks

had

tubcrous

sclerosis.

sharp

diffuse

waves.

compatible patients

with

activity “modified

had diffuse

Each patient

and

underwent

three

ding sessions from

~:OO a.m.

separated

weeks.

channels

by eight

two or four

and one channel respiration future

playback.

the polygraphic

Dr. King;

Polygraphic

data (EEG)

included

activity,

should

Department

onto the audio channel

Augusta,

GA 30912.

Received

April

record,

of

activity, and

polygraphic

of the video tape for

on the video

monitoring.

A trained

bc addressed

to:

10, 1985; accepted

two channels

the first eight channels

directly

of Neurology;

days

six or eight

electrocardiogram.

to the on-line

data were recorded

rccor-

separate

of clcctromyographic

By means of a rcformattcr,

the tape for continuous

and two

on three

each of clcctrooculogram,

data were multiplcxcd

[1,2.4,5,8,9];

with focal predominance. vidcolpolygraphic

(Fig 1.2). In addition

and

patterns

six-hour

channels

dassicaI

synchronous

to 3:00 p.m.

of clcctrocnccphalographic

accclcromctry,

with

had multifocal

variable

activity

each. Three

high amplitude

compatible

hypsarrhythmia”

cpilcptiform

Probable

“cpilcptiform”

had chaotic

five patients with

6.8

tonic-

mcningocnccphalitis,

in one patient

spikes

+

in three patients

documented

[1.2,4,5,8,9];

cpilcptiform

(10.2

seizures.

hypoxia

previous

Three patients

2 12.4 were fmt

had generalized

partial

cerebral

cnccphalopathy

in all ten patients.

(27.7

spasms

to 21 months

had no known causal factors. Intcrictal clcctrocnccphalograms waves,

infantile

simple

pcrinatal

Ages at rhc date

13 to 54 months

to spasms, rhrcc patients

factors included

trisomy-21.

from

The

from

In addition

Communications of

in the had not

each patient

six boys and four girls.

ranged

2 SD).

ranged

months).

and

included

into the study

“hypsarrhythmia”

Medical

of ACTH,

valproatc

All patients

entry into the study.

slow

From

[lo].

extensive evaluation to exclude progressive ncurologic discasc. Informed conscnt was obtained from the parents of each patient prior to

activity

Introduction Since the first description of infantile spasms [l], numerous authors have described the infantile spasm syndrome [2-81. Despite these descriptions, the nosologic boundariesof infantile spasmshave remained poorly defined [6]. Until recently, precise information concerning ictal phenomena has not been available [i’91. both deficiencies have hampered adequate evaluation of proposed treatments. Using a polygraphic /video detection system [9], Kellaway et al. first reported a quantitative study of infantile spasms[y]. The present researchwas designed

were the first ten patients of sodium

spasms

Prior to the study,

and post-immunization

King DW, Dyken PR, Spinks IL, Murvin AL. Infantile Spasms:Ictal phenomena. Pediat Neurol 1985;1:213-8.

study

into a larger study of the cffrcacy

Medical

technician

College

of

portion

of

was con-

of Georgia;

May 6, 1985.

Kinget al: InfantileSpasIns

213

Accelerometer R. Upper Arm

---

\;

I

‘6

1 -. c.‘.,I. -,/

EMG L. SCM

F&re

1.

Polygraphic

associated

with an EEG

that persisted

12)perskting

stantly

in attendance

entire

six hour

background defined

observing

wake-sleep

activity, as awake.

sleep were defined

hypnagogic Segments

state.

[-

Initially,

episodes

The myoclonic

The tonic phase was associated

with suppression

a log. The

Segments

activity.

motor

activity

Prolonged

consisting tonic

and correlation

by with

criteria

were

of a brief jerk with or without cpisodcs

were ex-

cluded; (2) Simultaneous (3) Simultaneous

EEG change; accclcromctcr

activation

and clcctromyographic

activity. It soon became clear that many patients activity activity. developed

without

a preceding

of arrest without

Since thcsc episodes had EEG characteristics spasms and appeared

rhcy were also included

214

had episodes of brief tonic

jerk or episodes

PEDIATRIC

similar

to bc pathophysiologically

RCdtS

motor to fully similar,

Seven of the ten patients demonstrated 27 clustersof

as spasms.

NEUROLOGY

1 I and

wake cycle A summary of the frequency of recorded spasmsand their relationship to the sleep-wake cycle is shown in Table 1. During the 180 hours of recording time, 1073 spasmswere recorded; 727 occurred during the 120.7 hours of awake recording time, and 150 occurred during the 57.1 hours of sleep. The frequency of spasms during wakefulness(7.7 spasms/hour) wasgreater than that during sleep (2.5 spasms/hour) (Wilcoxin matched pair, W = -7, p I .05). During the first 30 minutes after awakening, a time we refer to as the “critical” period, the frequency of spasmswas 13.1 /hour. This rate was not significantly different from the frequency during the remainder of wakefulness (6.2 spasms/hour).

were

episode was reviewed

and tonic-clonic

(Channels

[ - - - ] = tonic phase.

Frequency of spasmsand relationship to sleep-

as spasms:

(1) Bilateral tonic

was

of background

of awake

in stage II. III, IV, or REM

met each of the following

activity

contraction

of us to

and arousal

tbc ictal phenomena

which

Paper speed IO mm /sec. The spasm was

by a tonic contraction.

EEG changes. included

wak.efu/ness.

] = myoclonicphase,

by two

hypcrsynchrony.

of activity

followed

the end of the tonic phase of the spasm. Note the EMG

was reviewed

as sleep. Each video-taped

three of us to determine

wave complex”.

for spasms and keeping

events

spasm during

contraction

the tonicphase.

record

and

“sharp

beyond

through

polygraphic

ictal

of infant&

of two phases: a myoc/onic

activity

determine

record

composed

Vol.

1 No. 4

:

Accelerometer _____cr--L. Forearm Accelerometer R. Lower Leg n7yfv

’

*_cN__-

...

;‘.,

i

15op

‘,,.. .. ,: , Ifs1 ..

I 8

Figure associated

2.

Posjgrapbic

recording

with a slow wave complex

of infantile

spasm during by suppression

followed

wakefulness. of background

Paper speed 30 mm /sec. The onset of the spasm (arrow) activity

with superimposed

9 Hz rhythmic

activity

was

SQivntah’y

(brackets).

spasms (Table 1). Clusters were defined as 3 or more spasms occurring at a frequency of at least 1 per 30 seconds. Clusters were composed of 3 to 54 spasms (mean 17.4, median 12), the total duration of clusters varying from 27 seconds to as long as 10 minutes, 12 seconds. Within clusters, spasm frequency varied from 2 to 14/minute. The frequency of clusters during the “critical” period was 0.53 clusters /hour; during the

Table

1.

Temporal

Sleep-Wake

characteristics

Cycle

of 1.079 infantile

remainder of awake recording time it was 0.16 clusters/hour. No clusters occurred during sleep; however, there were six instances in which a brief spasm awakened the infant, immediately followed by a cluster during wakefulness. Phases Of the 1079 spasms recorded, 928 were available on tape and of sufficient technical quality for review of

spasms in 10 patients

No. of

No. of

No. of

SF-=

Cluscea

Spasms in

Recoding

Clusters Awake (Critical

Period)*

(Remainder

of

Wakefulness) Sleep

+critical TFrequency

503

120.9

(344

(14)

(253)

(26.2)

(587)

(15)

(250)

(94.7)

(6.4

refers to the fint 30 minutes

of spasms during

of spasms (spasms I hour)

29

1,079 period

w-1

929

150

Total

Frequency

Time

7.7t (13.1)

0

0

59.1

2.5t

29

503

180.0

6.0

after awakening.

the awake state was significantly

greater

than that during

sleep @ 5 .05).

King et al: Infantile

Spasms

215

Table 2.

Ictal phases of 928 infantile

spasms

No. of Phases Myoclonic

Patients alone

Myoclonic-tonic Myoclonic-arrest

Awake

Sleep

10

226

111

10

352

22

38

0

38

94 (10.1%)

2

Total 337 (36.3%) 374 (40.3 %)

Myoclonic-tonic-arrest

5

94

Tonic-alone

36 16

4

40

(4.3%)

Tonic-arrest

4 3

0

16

(1.7%)

Arrest-alone

3

29

0

29

(3.1%)

and trunk movements included both flexion and extension. Shoulder movement was most commonly flexor (i.e; the upper arm moved forward in relation to the trunk) and was often associatedwith adduction, abduction, or external rotation. Both extension and flexion occurred at the elbow. In the lower extremities, hip movements were most commonly flexor and adductor. In four patients, occasionalspasmswere characterized by a reversal from flexor to extensor movements during the change from the myoclonic to the tonic contraction. For example, one patient had spasmswhich began with a myoclonic contraction causing flexion at the elbow. After a brief pause, the arms were slowly raised above the head and extended at the elbows to a final position 180 ’ from the resting position. Reversalsin the lower extremities from initial flexion of the hips to extension of the hips and kneeswere noted in three patients. No spasmswere purely unilateral; however, asymmetry waspresent in 325 of the 887 spasmsin which it could be adequately evaluated. In four patients, the predominant side of motor activity wasconsistent from spasm to spasm; but in five patients, predominance shifted from side to side. In addition to right /left asymmetry, there was often a difference in intensity of motor contraction between the upper and lower extreities. There was primary involvement of the upper

ictal phenomena. Spasms were composed of one OI more of the following phases: a brief myoclonic contraction, a more prolonged tonic contraction, and /or an arrest of activity with no observable movement. The composition of spasms by phases is shown in Table 2. The most common types were myoclonic-tonic (Fig 1) and myoclonic alone. In all instancesin which two or more phases were present, the sequence of events proceeded in this order: myoclonic contraction, tonic contraction, arrest of activity. Of the 468 spasmswhich contained both myoclonic and tonic contractions, the tonic contraction represented a persistence of the myoclonic movement in 67.3 % ; in the remainder, the tonic contraction involved different muscle groups or was separated from the myoclonic phase by a brief pause. Postural Motor Activity Using previously described criteria [6,7], spasmswere classified on the basis of postural motor phenomena into four categories - flexor, extensor, mixed, and arrest. These results are shown in Table 3. All patients demonstrated “flexor” and “mixed” spasms, and eight patients demonstrated at leastthree types. Spasmswere composedof a broad spectrum of motor phenomena. Intensity varied from a slight shrug of the shouldersto massivejerks of the trunk and extremities. Facial muscle contraction was present in 42.0% of the 640 episodesin which the face could be observed. Neck Table 3.

Classitication

of 928 i&mile

extremities in 44.4% of the spasms, primary involvement of the lower extremities in 6.5 %I, and spasms by motor phenomena

No. of Types

Patients

Flexor

10

Extensor

216

PEDIATRIC

(4.1%)

0

NEUROLOGY

Awake

Sleep

282

104

Total 386 (41.6%)

6

146

5

151 (16.3%)

Mixed

10

334

28

362 (39.0%)

Arrest

3

29

0

Vol.

1 No. 4

29

(3.1%)

relatively equal involvement extremities in 49.0%.

of the upper and lower

ArsociatedPhenomena and Duration The eyelids were visible for study in 741 spasms. Of the 441 episodes in which the eyes were open at the onset, eye closure occurred during the spasm in 35. Of the 300 episodes in which the eyes were closed at the onset, eye opening occurred in 195. Eyelid fluttering was present during 67 spasms in seven patients. Respiratory change either during or following the spasm occurred in 77 spasms in six patients. Two patients made ictal grunting noises, and three patients had postictal gasping, sighing, or deep breathing. Nonspecific mouth or tongue movements were noted in six patients; automatisms involving extremities were not observed. Although crying did not occur during a spasm, it immediately followed 119 spasms in five patients. It was often associated with a cluster of spasms in which crying occurred between spasms, intermittently ceasing during the spasm itself. because the cessation of spasms was often gradual, the duration of spasms was at times arbitrarily determined. The duration varied from less than 1 to 28 seconds (mean 2.5 5 seconds, median 2.0 seconds). The majority of episodes (90.7%) were 5 seconds or less. Episodes composed solely of a myoclonic contraction were less than one second in duration. Those which included myoclonic and tonic phases or tonic alone varied from 1 to 10 seconds (mean 2.4 seconds). Those in which there was an arrest of activity either alone or following motor activity varied from 1 to 28 seconds (mean 5.8 seconds). The majority of infants resumed pre-ictal activity immediately following the spasm. Electrographic Charactetistics Of the 843 spasms in which a myoclonic contraction occurred, there was an associated paroxysmal electrographic event in 99.0% (Fig 1,2). An initial slow wave or slow complex was present in 65.8 % (Fig 2). A “sharp wave complex” composed of single or multiple sharp waves or spikes often with accompanying slow waves, was present in 33.1% (Fig 1). Tonic contractions and periods of arrest were usually associated with suppression of background activity, the period of suppression often persisting beyond the end of the clinical spasm (Fig 1). Five patients had 167 spasms (18%) in which there was rhythmic EEG activity varying in frequency from 8 to 16 Hz and in duration from 1 to 3 seconds (Fig 2). In

5 1 episodes, the rhythmic activity was superimposed on the initial slow or sharp wave. In the remainder, it followed the initial wave and was superimposed on the early period of suppression (Fig 2). In all instances in which both were present, suppression of activity persisted beyond the rhythmic activity. During the prolonged recording sessions, three patients demonstrated “subclinical” EEG seizures. These were composed of an initial slow or sharp wave complex followed by suppression of activity and superimposed rhythmic activity. During these events, no observable motor phenomena and no obvious arrest of activity were present. Discussion Kellaway et al. published the first video-tape analysis of infantile spasms [7]. The present study confirms these observations. The patients used in this analysis were participating in a long-term study of the efficacy of sodium valproate in the treatment of infantile spasms. As a criterion for admission, all patients had spasms which had remained refractory to standard antiepileptic drugs and ACTH. As a result, these patients were older and more likely to have significant neurologic dysfunction than patients in previous studies [2-5,7]. Previous authors have described the high frequency of infantile spasms [2,6], their tendency to occur in clusters [1,3-81, and their tendency to occur in the “twilight” state between wakefulness and sleep [ l,361. In the study of Kellaway et al., 97.5 % of spasms occurred during wakefulness, 2.5 % during sleep, and 78.3 O/6in clusters [7]. In the present group of refractory patients, 86% of spasms occurred during wakefulness, and the overall frequency of spasms was greater during wakefulness than during sleep. However, the percentage of spasms during sleep was greater than that reported in previous studies [7]. This finding may partially be secondary to more severe neurologic dysfunction in these .older, refractory patients. The clinical and EEG phenomena of spasms during sleep were similar to those seen during the awake state, and the episodes were not limited to the early stages of sleep. Thus, we think it is unlikely that these episodes represented ‘ ‘sleep starts ’ ’ or nocturnal myoclonus [11,12].

Approximately 50% of spasmsin this study occurred in clusters, the clusters occurring only during wakefulness. Although the small number of patients and high degree of variability precluded demonstration Kinget al: Infantilespasms

217

of statistical significance, the frequency of clusters during the frost 30 minutes after awakening was more than three times that of the frequency during the remainder of wakefulness. It is likely that additional data will show a significant difference between this “critical period’ ’ and the remainder of wakefulness. These data suggest that in the evaluation of proposed treatments for infantile spasms, appropriate controls should be included for differences in frequency which occur during wakefulness, sleep, and the period just after awakening. Infantile spasms have traditionally been classified on the basis of postural motor phenomena into three types: flexor, extensor, and mixed [3-8). Early studies emphasized that flexor spasms were most common [3-61; however, Kellaway et al. found mixed spasms to be most frequent [7]. In this study, 41.6% were classified flexor, 39.0% mixed, and 16.3% extensor. We think it probable that the differences among these studies are partially related to differences in criteria for classification and the difficulty in applying the criteria. In the present study, many episodes did not fit easily into one of these categories; no episodes were composed solely of flexion or extension of all muscle groups. In addition, the motor phenomena, and thus classification, were often dependent on factors such as the intensity of the spasm, the presence or absence of a myoclonic or tonic phase, and the position of the extremities prior to onset. Asymmetry of motor phenomena was observed more frequently than in previous studies (4,7]. Since our patients were older and likely had greater neurologic impairment than unselected patients with spasms, it is possible that age or degree of cerebral dysfunction is related to degree of asymmetry. In five patients the side of predominance shifted from spasm to spasm, suggesting that unilateral predominance is not always contralateral to maximal central nervous system damage. As noted by others [2,4,7], the electrographic pattern of an individual spasm commonly consists of an initial paroxysmal event associated with the myoclonic contraction and suppression of activity at the time of the tonic or arrest phase. In addition, 18.0% of the episodes in this study were associated with rhythmic ictal activity either superimposed upon or immediately following the initial complex. There was no correlation between the presence of rhythmic activity and clinical phenomena and, as previously noted [7], there was no correlation between electroencephalographic pattern

218

PEDIATRIC

NEUROLOGY

Vol.

1 No. 4

and classification into flexor, extensor, or mixed type of spasm. Despite the fact that the boundaries of the infantile spasm syndrome remain controversial, the present data suggest that for severely involved, refractory patients, there is a characteristic syndrome of clinical and electrographic phenomena. Spasms are composed of myoclonic, tonic, and arrest phenomena. The motor activity consists of a variety of movements and postures, both flexor and extensor. Spasms occur throughout the sleep-wake cycle, more frequently during wakefulness, and spasms often occur in clusters. Electrographically, spasms usually begin with an initial paroxysmal event often followed by suppression of activity with or without rhythmic activity. Knowledge of these clinical, electrographic, and temporal characteristics is important for correct diagnosis and for the evaluation of proposed treatments. WC thank

Diane

Batts.

the sodium

valproatc

manuscript.

This work

No. l-NS-6-2340,

RN for hct assistance in the coordination study

and Pam House

was supported

for preparation

in part by Research

from the National

Institutes

of of the

Contract,

of Health.

References [l]

West WJ. On a peculiar

form

of infantile

convulsions.

Lancct

1841;1:724-1. [2]

Gibbs

EL, Fleming

of hypsarhythmia [3]

DNC~IMII

childhood. [4]

Jeavons and

Medicine [S] Bruyn

Gibbs

Chao

PM,

Bower

a study

D.

Massive

PM,

BD.

of

Infantile

112 cases.

William

Bower

G. cds. Handbook 1974;15:

FA. Diagnosis

spasms. Pediatrics

and prognosis

1954;13:66-73.

spasms

in

infancy

Spasms:

a rcvicw

and

1955;4:61-72.

No. 15. London: Jeavons

Elscvicr, [6]

R,

Epilcpsia

lircraturc

MM,

and infantile

Clinics

in

Hcincman.

BD.

Infantile

of clinical

of the

Developmental

1964:82. spasms.

neurology.

In:

Vinkcn

New York:

P,

American

219-34.

Iacy JR, Pcnry JK. Infantile

spasms. New York:

Raven Press,

1976. [7]

KeUaway

characterization

P. Hrachovy

RA,

and quantification

Frost

JD,

of infantile

Zion

T.

spasms.

Precise

Ann Ncurol

1979;6:214-8. [8]

KeIIaway

Marsclli

P. Frost JD,

PL, Pippingcr

in pediatrics.

Hrachovy

RA.

Infantile

spasms.

CE, Pcnry JK. cds. Anticpilcptic

New York:

In:

drug therapy

Raven Press. 1983;115-36.

[9] Frost JD. Hrachovy RA, analysis and characterization

Kcllaway P. Zion T. Quantitative of infantile spasms. Epilcpsia

1978;19:273-82. [lo]

Dyken

effects ofvaIproatc [ll]

Oswald

PR, Durant

RI-I, Batts DG,

on infantile I. Sudden

King

spasms. Pcdia Ncurol bodily

jerks

in falling

DW.

Short

term

1985;1:34-7. asleep.

Brain

1959;82:92-103. [I21

Coleman

myoclonus) Sleeping

RM.

Periodic

and restless and waking

1982;265-95.

movements

legs syndrome.

disorders.

Menlo

in

sleep

In: Guillcminaulc Park,

(nocturnal C, cd.

CA: Addison-Wesley,