Pediat. Res.

3: 298-304 (1969) Epilepsy prematurity

febrile convulsions seizures

Studies on Convulsive Disorders in Young Children

I. Incidence of Febrile and Nonfebrile Convulsions by Age and Other Factors

BEAJ. VAN DEN BERG['^] and J.YERUSHALMY

The Child Health and Development Studies of the Division of Biostatistics, School of Public Health,
University of California, Berkeley, California;
the Permanente Medical Group and the Kaiser Foundation Research Institute, Oakland, California, USA

I n a cohort of 18,500 newborn infants, those who had experienced one or more convulsions were
studied. Data were available for the unaffected as well as for the affected children. All families were
members of a prepaid medical care program that provides comprehensive medical care.
By five years of age, two percent of the 18,500 children had had one or more febrile convulsions;
one percent had had nonfebrile convulsions. The incidence of the first nonfebrile seizure was highest
during the first month of life; febrile convulsions reached a high peak during the second year of life.
The risk of a febrile convulsion was equal for boys and girls but that of a nonfebrile seizure was
higher for boys. The nonfebrile group had an excess of children with birth weights between 1500 and
2500 g, especially those whose gestational age was 37 weeks or more. The distribution by birth weight
and gestational age of the febrile group was similar to that of the entire cohort.
Multiple episodes occurred in one-third of the children of the febrile group and in three-quarters
of the nonfebrile group. About 3 % of the children with febrile seizures later developed nonfebrile
convulsions; about 7% of the nonfebrile group had initially a febrile seizure. Nearly 31 % in the
nonfebrile group and 7 % in the febrile group had severe congenital anomalies; in the entire cohort
only 3 % had such an anomaly before they reached age five. Among the children in the nonfebrile
group were 10% with cerebral palsy and 16% with mental retardation including 5 % with both
disorders.

Speculation

As the cohort of 18,500 children grows older, a certain number of them will develop real epilepsy.
This study provides three distinct groups: children who had febrile convulsions, those who had non-
febrile convulsions, and those who had no seizures in early childhood. The future follow-up of these
groups will provide the necessary data for determining the relative rates of developing epilepsy asso-
ciated with the convulsion experience in early childhood.
 Introduction the obstetric and other specialty clinics of the Kaiser
Hospitals in the East Bay Area. The children were ob-
The unique features of these studies are first, that they served regularly in pediatric clinics. The records of
relate to all the children who experienced one or more visits to the out-patient pediatric clinics and to all
convulsions among a large cohort of newborn infants, other specialty clinics, as well as the records for any
and second, that data are available for the unaffected periods of hospitalization, are available in one com-
as well as for the affected children. I n the course of an prehensive file.
extensive investigation, in which a large number of The pediatrician could not often observe the child
pregnant women and their children were observed during a seizure; however, the child was nearly always
and followed longitudinally, all children who had ex- seen shortly after an episode. At that time great effort
perienced one or more convulsive seizures were identi- was exerted to obtain from the mother a detailed
fied and studied. Data on biologic, medical and envi- history and description of the attack. This enabled the
ronmental factors were obtained prospectively on all pediatrician to ascertain whether the episode was in-
the families of the cohort. Data were also available on deed a convulsive disorder. Since comprehensive in-
the course of pregnancy, labor, and delivery. and out-patient care was available for the entire family,
I n this setting, it is possible to provide fairly reliable it is unlikely that many definite convulsions were un-
incidence rates and to investigate the relation of con- reported.
vulsive disorders to the many factors that are available Episodes of convulsive disorders may occur con-
for the entire cohort. I n addition, information on treat- currently with other conditions, or the seizure may
ment and prognosis becomes available as the follow-up be the only observable phenomenon. Most frequently
of the children continues. It is hoped, therefore, that convulsions in the age groups studied occur concur-
these studies will provide data from which the dyna- rently with high fever and systemic infection. Such
mics of the different manifestations of these disorders convulsions are referred to as febrile. Convulsive dis-
may be explored. The present paper is limited to the orders that occur in the absence of fever are combined
provision of estimates of the incidence of febrile and under the heading of nonfebrile.
nonfebrile convulsions by age, sex, birth weight and Occurrences of convulsive disorders stimulated a
gestational age, and to the investigation of the possible number of diagnostic activities (table I). About one-
relation of convulsive disorders to severe congenital half of the children having nonfebrile seizures were
anomalies. hospitalized; more than one-half received out-patient
neurologic consultation in addition to consultations
during hospitalization; more than one-fourth had an
Material and Method x-ray examination of the skull; and nearly two-fifths
had an EEG examination. Even among the febrile
The cohort was composed of some 18,500 children group, nearly one-quarter was hospitalized and about
born alive between 1960 and 1967 in Kaiser Founda- one-quarter had a spinal fluid examination.
tion Hospital, Oakland, California. The parents were Fifty percent of the initial nonfebrile seizures were
members of the Kaiser Health Plan, a prepaid medical generalized myoclonic or the 'grand mal' type; twelve
care program in the San Francisco-East Bay area that
provides comprehensive medical in- and out-patient
services for the entire family. The women reported for Table I. Diagnostic procedures for children with febrile
prenatal care relatively early in their pregnancy. At and nonfebrile convulsive disorders
the first visit an extensive and comprehensive interview
was obtained. Participation of the pregnant women in Procedures Percent of children
the Child Health and Development Studies was nearly Febrile Nonfebrile
100 %.
The population was broadly based and typical of Hospitalizations 24.4
an employed population. It is deficient only in very Neurologic consultation1 5.7
indigent and very affluent segments. Socio-economi- X-ray skull 7.7
cally and educationally, the study population was of EEG 6.9
a somewhat higher level than that of the general San Echogram 0.8
Francisco-East Bay area. Ethnically, 65 % of the popu- Spinal fluid 28.0
lation were white, 23 % black, and the remainder were Ventriculogram -

-
of other or mixed ethnic backgrounds. Number of children 246
Information on illnesses and medication during
In out-patient clinic.
pregnancy was derived from the medical records of
300 VAN DEN BERG,YERUSHALMYStudies o n convulsive disorders i n young children.

percent were localized myoclonic; about the same per- Febrile convulsions were rare in the first few months of
cent were mainly hypertonic; and twenty percent were life, increased and reached a peak in the second year
akinetic episodes with loss of posture or spells ofstaring. of life, dropped sharply in the third year and slowly
Two percent of the children had typical symptoms and thereafter. Nonfebrile disorders, however, may occur
signs of 'petit mal' and three percent had 'infantile almost immediately after birth. Themonthly incidence
spasms'. Febrile seizures, however, were more uniform. of the first seizure of the nonfebrile disorders reached
Nearly ninety percent of the first febrile seizures were a peak in the first month of life and dropped sharply
generalized myoclonic and the remainder were mainly thereafter.
hypertonic or hypotonic spells. Figure 2 shows cumulative incidence rates. By the
age of five years, about two percent of the children had
Incidence of Convulsive Disorders
I n calculating the incidence rates, it is necessary to
take into consideration the fact that the length of
follow-up was not the same for all the children. The
youngest, born in 1967, have been followed for only
one year while the oldest, born in 1960, have been
followed for more than seven years. The average age
of the children, when last observed, was 3.2 years. The
i \ --- Febrile
incidence rates, therefore, were calculated on the basis
of the number of children who were actually observed
.' -Non-febrile

during each period.

The differences in incidence by ethnic group were
very small; therefore, the data are combined for all
groups.

Age a t j r s t seizure
0 :,I,~
1
Years
2 3 4 3!
l
Among the 18,500 children, 359 had experienced
one or more convulsions. The average monthly inci- Fig. I . Average monthly rates by age a t first seizure of
dence rate, by age at first attack, was dissimilar for febrile and nonfebrile convulsions.
febrile and nonfebrile disorders (table I1 and fig. 1). Child health and development studies, 1960-1967.

Table II. Incidence of febrile and nonfebrile convulsive disorders by age of first seizure. Child health and
development studies 1960-1 967

Children with Average monthly Cumulative Incidence

convulsions incidence rates per 1000 rates per 1000
F1 NF2 Total F NF Total F NF Total

0-i 1 month
1-i 3 months
3-< 6 months
6-i 9 months
9-i 12 months
1-i 2 years
2-< 3 years
3-1 4 years
4-i 5 years
5-i 6 years

Total 246 113 359

F = febrile convulsion
NF = nonfebrile convulsion
 I. Incidence of febrile a n d nonfebrile convulsions b y age a n d other factors 30 1

experienced one or more febrile convulsions and one febrile children, 6.4 % were in this birth weight-gesta-
percent had a nonfebrile convulsion. tion group, while only 2.5 % of the febrile children and
Since infections of the respiratory tract and otitis 2.3 % of the total cohort were in Group 111. The chil-
media, accompanied by high fever, are common in dren of birth weight-gestation Group I11 have been
young children, it was not surprising that in the present reported to have also a larger number of severe con-
study about 63 % of the febrile convulsions were asso- genital anomalies [14] ; this fact, however, does not
ciated with these conditions. Thirteen percent were explain entirely their high proportion of nonfebrile
associated with roseola, and the remainder were asso- convulsions, because the excess is noted also when the
ciated with a variety of infectious diseases. Ten children children with severe anomalies were excluded.
or 4.0 % of those with febrile convulsions had intra-
cranial infections. Multible Etisodes
There was a wide range in the degree of severity of
Sex convulsive disorders, varying from a single almost acci-
I n the febrile group 131 (53.3 %) were boys and 115 dental episode to very severe daily attacks. It was de-
(46.7 %) were girls. These percentages were not signifi-

I
cantly different from those of the total cohort (51.1 %
35 Total -
boys and 48.9 % girls). Among the nonfebrile group,
Febrile
however, there was a statistically significant excess of 301 Non-febrile ----
boys (61.9 % boys; p <0.05).

Birth Weight and Gestational Age

Table I11 compares the distribution of the febrile
and nonfebrile groups by birth weight and gestational
age with that of the entire cohort of liveborn infants
surviving the first month of life. The distribution of the
febrile group by birth weight and gestational age was
similar to that of the entire cohort. The nonfebrile
group, however, had an excess of children with birth
weights between 1500 and 2500 g (Groups I1 and 111)
compared with the entire cohort (10.0 % compared
with 4.6 %; p < 0.05).
I 1
Ege in years
2 3 4 5
I
The excess was most pronounced in Group III- Fig. 2. Cumulative rates by age a t first seizure of febrile
composed of children with birth weights of 1500-2500 and nonfebrile convulsions.
g and gestational ages of 37 or more weeks. Of the non- Child health and development studies, 1960-1967.

Table ZIZ. Percent distribution in 5 birth weight-gestation groups of children having febrile and nonfebrile con-
vulsions. Child health and development studies, 1960-1967

Group Birth weight Gestational age Febrile Nonfebrile Cohort1

(9) (weeks) 1960-1967

5 1500 all gestations 0.4 - 0.25

1501-2500 < 37 1.6 3.6 2.25
1501-2500 37+ 2.5 6.4 2.3
> 2500 < 37
+
4.5 5.5 5.1
> 2500 37 91.0 84.5 90.1

Total percent
I-V number

1 Surviving the first month of life.

2 Excludes children for whom the birth weight-gestation group was not known.
302 VAN DEN BERG,YERUSHALMYStudies o n convulsive disorders i n young children.

sirable, therefore, to differentiate between infants hav- and congenital heart disease. Moderate anomalies re-
ing multiple seizures and those with only a single at- presented the intermediate category and included
tack. Multiple episodes occurred in one-third of the conditions such as esotropia and exotropia, ptosis pal-
children in the febrile group and in three-quarters of pebrae, first degree hypospadias, polydactyly, and
those in the nonfebrile group. In a large proportion congenital dislocation of the hip. Trivial anomalies
of the children who experienced multiple seizures, the interfered only slightly, temporarily, or not at all with
second attack followed within less than a year of the the child's health, and included metatarsus varus, skin
first. The chance of the child having a second attack tags, and blocked tear duct.
was independent of the child's age at which the first In the present paper, only the category of severe
episode occurred. Of the 254 children who had an anomalies is reported.
initial febrile seizure, eight or 3.1 % later developed A large proportion of the children in the nonfebrile
nonfebrile convulsions. It can also be stated that eight group had severe congenital anomalies, especially
or 7.1 % of the 113 children with nonfebrile convulsions anomalies of the central nervous system. Thirty-one
had an initial febrile seizure. percent of the children in the nonfebrile group and 7 %
The frequency of special diagnostic procedures in- of those in the febrile group had severe congenital
creased greatly after the occurrence of a second con- anomalies (table IV). In comparison, only 3 % of the
vulsion; ninety-five percent of the children who experi- children of the entire cohort who survived the first
enced multiple episodes of convulsions were given an month of life suffered a severe congenital anomaly by
EEG examination. age five. Thus, even among the febrile group, the fre-
quency of severe congenital anomalies was significantly
higher than in the entire cohort (p i0.05).
Severe Congenital Anomalies In the febrile group, the frequency of severe congen-
ital anomalies of the central nervous system was 1.2 %,
The procedure adopted by the Child Health and De- a rate coming very close to that of the entire cohort. I n
velopment Studies in regard to congenital anomalies the nonfebrile group, however, 30 children or 26.5 %
was first to ascertain the specific diagnostic entity in had such anomalies; seven of these children also had
the necessary detail. For many of the analyses, how- severe anomalies of other systems.
ever, the specific diagnostic entities were combined The frequency of severe congenital anomalies was
and investigated in three categories according to the much higher in children who experienced multiple
severity of the anomaly: severe, moderate, and trivial. episodes of seizures than in those who experienced only
Some examples may clarify the categories. Severe a single episode; however, even the latter had higher
anomalies included conditions that, if not corrected, frequencies than is true for the entire cohort.
would interfere seriously with the child's well being In the nonfebrile group, the frequency of cerebral
and/or development or were potentially hazardous. palsy and of mental retardation was extremely high
Examples are hypothyroidism, microcephaly, cleft lip, (table IV). Eleven or 9.7 % of the children in the non-

Table IV. Number and percent of severe congenital anomalies in children having single or multiple episodes of
febrile and nonfebrile convulsions

Convulsive disorder Number of Severe congenital anomalies

children Total of central nervous system
Total Cerebral palsy Mental
retardation
No. % No. % No. % No. %

Febrile total
1 episode
2 or more episodes
Nonfebrile total
1 episode
2 or more episodes

Include 6 who had both cerebral palsy and mental retardation.

 I. Incidence of febrile a n d nonfebrile convulsions by age a n d other factors 303

febrile group had cerebral palsy and 18 or 15.9 % were phenomena, some of which confirmed previous esti-
mentally retarded. These figures include 6 children mates derived clinically, and others of which have not
who had cerebral palsy as well as mental retardation. been reported. The derived incidence rate of 3 % of
Thus, 23, or one-fifth of the 113 children in the non- convulsive disorders by five years of age confirmed
febrile group, had cerebral palsy or mental retardation estimates obtained clinically by other investigators [4,
or both. This figure may be compared with 1.2 % in 7, 111. More refined estimates of the incidence rate by
the febrile group and 0.5 % in the cohort. The number age at the time of the first seizure of febrile and non-
of mentally retarded children is understated sincemany febrile disorders are provided.
of the children were still too young for a reliable esti- The peak risk of a first convulsion in the nonfebrile
mate of their mental development. group occurred during the first month of life, while that
The relation of anomalies of the central nervous for the febrile group occurred during the second year
system to febrile and nonfebrile convulsions may be of life. Since the incidence of acute infections has been
considered in another way. In the cohort there were a found to be as high during the first as it is during the
total of 165 children with severe anomalies of the cen- second year of life [13], it is suggested that the risk of
tral nervous system. In these children, the incidence of febrile convulsions is not entirely a function of the fre-
febrile convulsions was the same as that of the total quency of acute infections.
cohort (1.8 %), while the incidence of nonfebrile con- The method used in the present study provides
vulsions was ten times as high (18.2 %). quantitative expressions for the risk of repeated at-
Among the 165 children there were 34 with cerebral tacks. The percentages of recurrence of seizures re-
palsy and 62 with mental retardation, including 12 ported by other investigators [2, 6, 9, 121 vary widely,
with both conditions. Thus, the incidence of a non- and are, in general, higher than those of this study.
febrile convulsion was 32.3 % for children with cerebral Quantitative expressions were derived also for the
palsy, 29.0 % for mentally retarded children, and relation of convulsive disorders to congenital anoma-
50.0 % for children with both conditions. lies, anomalies of the central nervous system, cerebral
palsy
- . and mental retardation. While this relation has
Mortality been commented on before, no specific figures have
There was a relatively high mortality of children been reported [3, 101.
with convulsive disorders, especially of those in the In the present study, 3 % of the children who had
nonfebrile group. There were four deaths among the an initial febrile seizure had later nonfebrile con-
246 children with febrile disorders and eight deaths vulsions. This percentage is of the same order of magni-
among the 113 children with nonfebrile disorders. tude as that reported as 'simple febrile disorders' by
Since all children had passed the neonatal period, this LIVINGSTON [9] and those reported by Friderichsen
number of deaths exceeded the expected number (4 and MELCHIOR [5] and by MILLICHAP [12].
versus 1 1/, after the first month of life in the febrile The role of prematurity in convulsive disorders can
group and 8 versus less than 1 in the nonfebrile group). be adequately studied when the distribution by birth
This excess does not necessarily reflect an extra risk of weight and gestational age is compared with that of
death associated with convulsions per se, but may re- the entire cohort. Febrile convulsions appeared as fre-
flect the high level of mortality of the associated con- quently among full term infants of 'normal' birth
ditions. For example, in the nonfebrile group, the weight as among children who were born preterm
eight children who died had severe congenital anom- or with low birth weights. Nonfebrile convulsions,
alies. however, appeared twice as frequently among infants
I n the febrile group, two children died from the of low birth weight; a similar excess in the incidence of
infection that triggered the initial seizure, one from a low birth weight has been reported for children with
general infection of unknown origin and one from viral epilepsy [8]. Our data indicate that the risk of non-
encephalitis. The third child died of complications of febrile seizures is especially high for infants of low birth
a congenital anomaly (biliary atresia) and the fourth weight and gestational age of 37 or more weeks. It is
child had a fatal accident. important to observe, however, that more than 85 %
of the children with nonfebrile convulsions were 'ma-
ture' by both birth weight and gestational age criteria.
Discussion

This study provided the opportunity to observe and Summary

study children with convulsive disorders in the setting
of an entire cohort of which they were a part. It was This study is based on all children ascertained as having
possible to obtain quantitative expressions for certain had convulsive disorders in a cohort of some 18,500
children; 246 children had febrile convulsions and 113 5. FRIDERICHSEN, C. and MELCHIOR, J.: Febrile con-
had nonfebrile convulsions. The risk of a convulsive vulsions in children, their frequency and prognosis.
disorder by five years of age was found to be 2 % and Acta pediat. Scand. 43: (Suppl. 100) 307-317
1 % for febrile and nonfebrile seizures respectively. (1954).
The highest incidence of the initial seizure for children 6. LENNOX, W. G. : Significance of febrile convulsions.
of the febrile group occurred during the second year of Pediatrics 11: 341-357 (1953).
life, and for children of the nonfebrile group during the 7. LENNOX, W. G. : Epilepsy and related disorders,
first month of life. Nearly one-third of the children in vol. 1 (Little, Brown, Boston/Toronto 1960).
the febrile group and two-thirds of the children in the 8. LILIENFELD, A. M. and PASAMANICK, B. : Associa-
nonfebrile group had multiple episodes of seizures. tion of maternal and fetal factors with the develop-
About 3 % of the children having initial febrile seizures ment of epilepsy. 1.Abnormaliiies in the prenatal
later had nonfebrile convulsions. and paranatal periods. J. amer. med. Ass. 155:
The risk of febrile seizures was about equal for boys 719-724 (1954).
and girls, but relatively more boys than girls suffered 9. LIVINGSTON, S.: Convulsive disorders in infants
nonfebrile seizures. In the febrile group, the relative and children. Advanc. Pediat., vol. 10, pp. 113-195
frequency of low birth weight children was similar to (Year Book Publishers, Chicago, Ill. 1958).
that of the total cohort, but the nonfebrile group had 10. MASON,T. and NIEDERMEYER, E.: Convulsive
an excess of low birth weight children, especially those disorders during the first three months of life. Epi-
whose gestational age was 37 or more weeks. The fre- lepsia 9: 1-9 (1968).
quency of severe congenital anomalies was 7 % in the 11. MILLICHAP, J. G. : Diagnosis and management of
febrile group and 31 % in the nonfebrile group, com- convulsive disorders with special emphasis on sei-
pared with 3 % for the entire cohort. The majority of zures amenable to specific therapies. Pediat. Clin.
severe anomalies in the nonfebrile group were those N. Amer. 7: 583-608 (1960).
of the central nervous system, especially manifested as 12. MILLICHAP, J. G. : Febrile convulsions (Macmillan,
mental retardation and cerebral palsy. New York 1968).
13. VAN DEN BERG, B. J.: Morbidity of low birth
weight and/or preterm children compared to that
of the 'Mature'. I. Methodological considerations,
Rcfirences und Notes and findings for the first two years of life. Pediatrics
42: 590-597 (1968).
1. BREG,W. R. and YANNET, H. : The child in a con- 14. YERUSHALMY, J. ; VAN DEN BERG,B.J. ; ERHARDT,
vulsion. Pediat. Clin. N. Amer. 9: 101-1 12 (1962). C. L. and JACOBZINER, H. : Birth weight and gesta-
2. CARTER,S.: Diagnosis and treatment: Manage- tion as indices of 'Immaturity' neonatal mortality
ment of the child who has had one convulsion. and congenital anomalies of the 'immature'. Amer.
Pediatrics 33: 43 1-434 (1964). J. Dis. Child. 109: 43-57 (1965).
3. DENNHOFF, E. and ROBINAULT, I. P. : Cerebral 15. Supported by PHs Grant No. HD-00718 of the
palsy and related disorders, a developmental ap- National Institutes of Health, U.S. Public Health
proach to dysfunction (McGraw-Hill, New York/ Service.
Toronto/London 1960). 16. Requests for reprints should be addressed to: Dr.
4. FARMER, T.W. : In Pediatric Neurology (ed. FAR- BEAJ. VAN DEN BERG,Child Health and Develop-
MER, T H . ~ . )chapter
, 2, pp.65-108 (Harper and ment Studies, University of California, Berkeley
Row, New York 1964). 3867 Howe St., Oakland, Cal. 9461 1, USA.