Brain & Development xxx (2017) xxx–xxx

www.elsevier.com/locate/braindev

Original article

Predictive value of EEG for febrile seizure recurrence

Alberto M. Cappellari a,⇑, Carolina Brizio b, Marta B. Mazzoni b, Giuseppe Bertolozzi b,
Federica Vianello b, Alessia Rocchi b, Massimo Belli a, Andrea Nossa a, Dario Consonni c,
Gregorio P. Milani d, Emilio F. Fossali b
a
Department of Neuroscience, Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico, Milan, Italy
b
Department of Pediatric Emergency, Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico, Clinica De Marchi, Milan, Italy
c
Epidemiology Unit, Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico, Milan, Italy
d
Pediatric Unit, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico and Department of Clinical Sciences and Community
Health, Universita‘ degli Studi di Milano, Milan, Italy

Received 10 October 2017; received in revised form 2 December 2017; accepted 5 December 2017

Abstract

Objective: To define the role of the EEG in predicting recurrence of febrile seizures (FS) in children after a first FS.
Methods: Children with a first simple or complex FS who underwent EEG at our hospital were retrospectively enrolled. EEG
recordings were classified in three groups: normal, abnormal (slow activity or epileptiform discharges), and pseudo-petit mal dis-
charge (PPMD) pattern. Children were followed-up for at least three years.
Results: A total of 126 patients met the entry criteria, and 113 of them completed the follow-up. Risk of FS recurrence decreased
linearly with increasing age ( 2% per month). The risk was higher among patients with PPMD pattern (absolute risk 86%, adjusted
relative risk 2.00) and abnormal EEG (epileptiform discharges: absolute risk 71%, adjusted relative risk 2.00; slow activity: absolute
risk 56%, adjusted relative risk 1.44), compared with those with normal EEG (absolute risk 41%).
Conclusions: PPMD and abnormal EEG should be considered as an independent risk factor for FS recurrence.
Ó 2017 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.

Keywords: Febrile seizures; EEG; Pseudo-petit mal discharge

1. Introduction normal neurodevelopment [3], while the value of EEG

Febrile seizures (FS) are the most common type of
seizures in childhood [1]. Most of them are simple FS,
while 25–35% are complex FS [2]. The American
Academy of Pediatrics guidelines state that electroen-
cephalogram (EEG) is not justified in the evaluation of
a neurologically healthy child with a simple FS and
⇑ Corresponding author at: Fondazione IRCCS Ca’ Granda,
Ospedale Maggiore Policlinico, Via Sforza 35, 20122 Milan, Italy.
E-mail address: alberto.cappellari@policlinico.mi.it
Cappellari).
in patients with complex FS remains controversial [4].
However, patients with simple FS are occasionally
investigated by EEG because of parental demand, FS
recurrence or their physician’s recommendation [5]. Fur-
thermore, even in children with simple FS, abnormal
EEG findings may increase the risk of subsequent
unprovoked seizures [6]. There is longstanding debate
on the usefulness of EEG in children with FS to predict
long-term outcome [7], and some authors reported that
there is no evidence that EEG abnormalities help to pre-
(A.M. dict either the recurrence of FS or the development of
epilepsy [8,9]. However, the predictive value of EEG

https://doi.org/10.1016/j.braindev.2017.12.004
0387-7604/Ó 2017 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Cappellari AM et al. Predictive value of EEG for febrile seizure recurrence. Brain Dev (2017), https://doi.org/
10.1016/j.braindev.2017.12.004
2 A.M. Cappellari et al. / Brain & Development xxx (2017) xxx–xxx
patterns different from epileptiform discharges has not
been fully investigated. The main aim of our study was
to evaluate the possible role of EEG in predicting recur-
rence of FS in children after a first FS, during a follow-
up of at least three years.
2. Patients and methods
2.1. Study design
This was a retrospective chart review of consecutive
children with a first FS who underwent EEG at our hos-
pital from January 2007 to January 2013. Informed con-
sent was obtained from parents of all patients.
2.2. Patient population
We included children aged 6–72 months evaluated in
our hospital after a first FS, who had both an EEG and
a follow-up of at least three years. Although FS is usu-
ally defined to occur in infants or children 6–60 months
of age, we included infants until 72 months owing to
possible occurrence of simple FS in Genetic Epilepsy
with Febrile Seizures Plus (GEFS+), whose onset is typ-
ically between 6 months and 6 years of age [10]. Chil-
dren with previous unprovoked seizures or known
neurological disorders (two patients with cerebral palsy)
were excluded from the study.
2.3. Definitions
FS was defined as a seizure occurring in association
with a body temperature 38 °C in the absence of cen-
tral nervous system infection, metabolic disturbance,
or acute electrolyte imbalance. Simple FS were defined
as generalized, with duration <15 min and not recurrent
within 24 h. Complex FS were either focal, repetitive
within 24 h or with a duration >15 min. Recurrent FS
was defined as a seizure occurring during a new episode
of fever defined as above in a child who had already
experienced at least one previous FS. Epilepsy was
defined as two or more unprovoked seizures occurring
>24 h apart.
2.4. Procedure
EEG studies were performed using the 10–20 interna-
tional system with bipolar montages, after a median
time interval of 30 days (range 10–60). EEG was
obtained without pharmacological sedation. Each
recording included wake and sleep stages, and activation
procedures were performed when possible. EEG was
interpreted by a pediatric neurologist and neurophysiol-
ogist (AMC). EEG findings were classified in three
groups: (1) normal: background activity appropriate
Please cite this article in press as: Cappellari AM et al. Predictive value of EEG for febrile seizure recurrence. Brain Dev (2017), https://doi.org/
10.1016/j.braindev.2017.12.004
for age, without any abnormality, (2) abnormal: focal
or generalized slow activity, or epileptiform discharges,
(3) PPMD pattern: paroxysmal nearly generalized high
voltage 3–4/sec waves with poorly developed spikes
between the slow waves, occurring in drowsiness [11]
(Fig. 1). Information on FS recurrence and occurrence
of unprovoked seizures was obtained by a telephone
interview, during a follow-up of at least 3 years.
2.5. Outcome measures
Main outcome measure was the recurrence of febrile
seizures (FS) in children with a first FS.
2.6. Statistical analysis
Data were analyzed by chi-squared test for categori-
cal variables and Wilcoxon-Mann-Whitney test for the
continuous ones. Univariate and multiple Poisson
regression models with robust variance were fitted to
calculate relative risks (RR) of FS recurrence and 95%
confidence intervals (95% CI) [12]. Statistical analyses
were performed using Stata 13 (StataCorp. 2013. Stata:
Release 13. Statistical Software. College Station, TX:
StataCorp LP).
3. Results
A total of 126 patients met the entry criteria, and
113 of them completed the follow-up. Clinical and
EEG characteristics at admission of 113 children with
complete follow-up are illustrated in Table 1, and their
risks of recurrence of FS are showed in Table 2. First
seizure duration varied from 0.5 to 40 min (median 2,
mean 3.5), 6 patients presented with seizure >15 min,
while data on the second seizure duration were not
available in all patients. Risk of FS recurrence
decreased linearly with increasing age ( 2% per
month). Mean age at the time of recurrence of FS
was 30 months. Time between the onset of fever and
the initial seizure was <1-h only in 4 children. However,
we lacked data on 25 children. In any case, we verified
that risk did not vary much (RR = 1.01 per hour). For
these reasons we did not include this variable in the
final regression model. Children with PPMD and
abnormal EEG pattern showed a higher risk of FS
recurrence (PPMD: absolute risk 86%, adjusted relative
risk 2.00; epileptiform discharges: absolute risk 71%,
adjusted relative risk 2.00; slow activity: absolute risk
56%, adjusted relative risk 1.44) compared to children
with normal EEG (absolute risk 41%). Children
with PPMD and epileptiform discharges had more
frequently a family history of FS than children with
normal EEG, about three times, but we evaluated risk
of FS adjusted for family history of FS.
 A.M. Cappellari et al. / Brain & Development xxx (2017) xxx–xxx 3

Fig. 1. Pseudo-petit mal discharge: paroxysmal generalized high voltage slow-waves with intermixed poorly developed spikes, occurring in
drowsiness.

Table 1
Clinical and EEG characteristics at admission in 113 children with complete follow-up.
Variable All EEG
Normal PPMD Slow Epileptic
Patients 113 68 (100) 29 (100) 9 (100) 7 (100)
Age (months) 24 [8–71] 23.5 [8–71] 26 [8–62] 31 [11–70] 28 [19–62]
Gender (F/M) 51/62 31/37 12/17 4/5 4/3
Type of FS
Simple 87 (77) 52 (76) 26 (89) 5 (56) 4 (57)
Complex 26 (23) 16 (24) 3 (11) 4 (44) 3 (43)
Focal* 10 (38) 5 (31) 2 (67) 1 (25) 2 (67)
Repetitive within 24 h* 10 (38) 8 (50) 1 (33) 1 (25) 0
Duration > 15 min* 6 (23) 3 (19) 0 2 (50) 1 (33)
Family history of FS 20 (18) 8 (12) 9 (31) 1 (11) 2 (29)
Family history of epilepsy 6 (5) 3 (4) 2 (7) 0 (0) 1 (14)
Maximal axillary temperature
<39 °C 52 (46) 35 (51) 9 (31) 5 (56) 3 (43)
39–39.9*C 45 (40) 22 (32) 15 (52) 4 (44) 4 (57)
40* C 16 (14) 11 (16) 5 (17) 0 (0) 0 (0)
Abbreviations: FS, febrile seizures; PPMD, pseudo-petit mal discharge.
Results are reported as median [range], or as number (%).
*
Percentages calculated over children with complex febrile seizures.

Please cite this article in press as: Cappellari AM et al. Predictive value of EEG for febrile seizure recurrence. Brain Dev (2017), https://doi.org/
10.1016/j.braindev.2017.12.004
4 A.M. Cappellari et al. / Brain & Development xxx (2017) xxx–xxx

Table 2
Absolute and relative risks (RR) and 95% confidence intervals (CI) of recurrence of febrile seizures according to selected characteristics, in 113
children with complete follow-up.
Variable FS recurrence N (%) RR, crude 95% CI RR, adjusted* 95% CI
Age, months 63 (56) 0.98 0.97–99 0.98 0.97–0.99
Gender
Female 28 (55) 1.00 Reference 1.00 Reference
Male 35 (56) 1.03 0.74–1.43 1.13 0.81–1.56
Family history of FS
No 49 (53) 1.00 Reference 1.00 Reference
Yes 14 (70) 1.33 0.94–1.88 1.06 0.73–1.54
Family history of epilepsy
No 59 (55) 1.00 Reference 1.00 Reference
Yes 4 (67) 1.21 0.67–2.19 1.06 0.61–1.84
Type of FS
Simple 49 (56) 1.00 Reference 1.00 Reference
Complex 14 (54) 0.96 0.64–1.43 0.92 0.62–1.37
Maximal axillary temperature
<39 °C 27 (52) 1.00 Reference 1.00 Reference
39–39.9 °C 28 (62) 1.20 0.85–1.70 1.03 0.74–1.44
40 °C 8 (50) 0.96 0.55–1.68 0.88 0.53–1.47
EEG pattern
Normal 28 (41) 1.00 Reference 1.00 Reference
PPMD 25 (86) 2.09 1.52–2.88 2.00 1.42–2.82
Slow 5 (56) 1.35 0.70–2.59 1.44 0.80–2.57
Epileptiform 5 (71) 1.73 1.00–3.01 2.00 1.20–3.32
Abbreviations: FS, febrile seizures; PPMD, pseudo-petit mal discharge.
*
Each variable adjusted for the others.

4. Discussion 43%, slow activity 44%). Furthermore, PPMD was never

Our study indicates that PPMD pattern and abnor-
mal EEG are predictive of FS recurrence. In 1964,
Gibbs and Gibbs described PPMD as a paroxysmal dis-
charge consisting of generalized or nearly generalized
high voltage 3–4/sec waves with a poorly developed
spike in the positive trough between the slow waves
occurring in drowsiness only. A history of FS was quite
common in these children [11]. In 1983, Alvarez et al.
described a similar pattern, which they called ‘‘hypna-
gogic paroxysmal spike and wave activity”. This finding
was found in 23% of patients with FS, while it was not
observed in the normal control group [13]. However, the
usefulness of PPMD as a predictive marker for recur-
rence of FS has not been subsequently investigated. In
our study, the incidence of PPMD in children with FS
(25.7%) was similar to that reported by Alvarez et al.
[13]. Furthermore, absolute risk of FS recurrence was
higher among those with PPMD (86%) and abnormal
EEG (epileptiform discharges: 71%; slow activity:
56%), compared with those with normal EEG (41%).
Patients with PPMD and epileptiform discharges could
present with simple or complex FS. Interestingly,
PPMD had a very low incidence of complex FS (11%)
compared with abnormal EEG (epileptiform discharges
detected in the subgroup of complex FS with duration
>15 min. We found no relationship between the location
of epileptiform discharges and recurrence of FS. Type of
recurrent FS, simple or complex, was not fully investi-
gated at follow-up. Therefore, we cannot make a rela-
tionship between EEG at admission and type of
recurrent FS.
In our population, 41% of patients with normal EEG
had FS recurrence. We run the regression models among
children with normal EEG: apart from age we did not
find other important risk factors of FS recurrence. Pos-
sibly, our patients represent a selected sample admitted
to a tertiary care hospital. Some authors reported no sig-
nificant correlation between the recurrence rate of FS in
patients with normal EEG and that of patients whose
EEGs showed epileptiform discharges [5].
Clinical predictors of recurrent FS are well-known
since the work by Berg et al., consisting of young age
at onset, history of FS in a first-degree relative, low
degree of fever while in the emergency department and
brief duration (<1-h) of recognized fever prior to the ini-
tial seizure [14]. As reported by the authors, age of onset
was examined both as a simple dichotomous factor (<18
vs  18 month) and as an ordinal variable measured in
6-month increments after the age of 18 months [14].

Please cite this article in press as: Cappellari AM et al. Predictive value of EEG for febrile seizure recurrence. Brain Dev (2017), https://doi.org/
10.1016/j.braindev.2017.12.004
 A.M. Cappellari et al. / Brain & Development xxx (2017) xxx–xxx
We also found that younger age was a risk factor for FS
recurrence, and the risk decreased linearly with increas-
ing age (-2% per month). Nonetheless, multivariate anal-
ysis identified PPMD as an independent risk factor of
FS recurrence.
Our study has several limitations, including its retro-
spective nature, short duration of follow-up and timing
of EEG. The time elapsed from FS is one of several vari-
ables reported to explain the wide range (2–86%) of
EEG abnormalities in FS [15]. However, a recent study
suggests that distribution of abnormalities in early EEG
(24–48 h) and late EEG (2 weeks) recording was the
same in patients with FS [16].
There is a debate on the role of EEG in children with
FS to predict long-term outcome [7]. Some authors sug-
gest that epileptiform discharges on EEG are not predic-
tive factors of the recurrence of FS, but important
predictors of epilepsy development [5,6]. However, to
our knowledge, the predictive value of other EEG find-
ings such as PPMD has not been fully investigated in the
literature.
FS recurrence is an important issue for both par-
ents and clinicians because it is a frightening experi-
ence. Young age at onset, family history of FS, low
degree of fever and short interval between fever onset
and initial seizure has been reported as main predic-
tors of recurrent FS [14]. Our study indicates that
PPMD pattern and abnormal EEG should be
considered as another independent risk factor for FS
recurrence.
5. Conclusion
EEG may have a role in predicting recurrence of FS
in children after a first FS. Although the limitations of
our study, we suggest that the practice of obtaining
EEG after a first FS needs better clarification by
prospective studies.
Conflicts of interest
None.
Financial support
None.
This research did not receive any specific grant from
funding agencies in the public, commercial or not-for-
profit sectors.
Please cite this article in press as: Cappellari AM et al. Predictive value of EEG for febrile seizure recurrence. Brain Dev (2017), https://doi.org/
10.1016/j.braindev.2017.12.004
5
References
[1] Carapetian S, Hageman J, Lyons E, Leonard D, Janies K, Kelley
K, Fuchs S. Emergency department evaluation and management
of children with simple febrile seizures. Clin Pediatr 2015;54(10):
992–8.
[2] Harini C, Nagarajan E, Kimia AA, de Carvalho RM, An S,
Bergin AM, Takeoka M, Pearl PL, Loddenkemper T. Utility of
initial EEG in first complex febrile seizure. Epilepsy Behav
2015;52:200–4.
[3] American Academy of Pediatrics (AAP). Febrile seizures: guide-
line for the neurodiagnostic evaluation of the child with a simple
febrile seizure. Subcommittee on febrile seizures. Pediatrics
2011;127:389–94.
[4] Kim H, Byun SH, Kim JS, Lim BC, Chae JH, Choi J, Kim KJ,
Hwang YS, Hwang H. Clinical and EEG risk factors for
subsequent epilepsy in patients with complex febrile seizures.
Epilepsy Res 2013;105:158–63.
[5] Wo SB, Lee JH, Lee YJ, Sung TJ, Lee KH, Kim SK. Risk for
developing epilepsy and epileptiform discharges on EEG in
patients with febrile seizures. Brain Dev 2013;35:307–11.
[6] Hwang G, Kang HS, Park SY, Han KH, Kim SH. Predictors of
unprovoked seizure after febrile seizure: short-term outcomes.
Brain Dev 2015;37:315–21.
[7] Natsume J, Hamano SI, Iyoda K, Kanemura H, Kubota M,
Mimaki M, Niijima S, Tanabe T, Yoshinaga H, Kojimahara N,
Komaki H, Sugai K, Fukuda T, Maegaki Y, Sugie H. New
guidelines for management of febrile seizures in Japan. Brain Dev
2017;39:2–9.
[8] Kuturec M, Emoto SE, Sofijanov N, Dukovski M, Duma F,
Ellenberg JH, Hirtz DG, Nelson KB. Febrile seizures: is the EEG
a useful predictor of recurrences? Clin Pediatr 1997;36:31–6.
[9] Pavlidou E, Panteliadis C. Prognostic factors for subsequent
epilepsy in children with febrile seizures. Epilepsia
2013;54:2101–7.
[10] Wirrell EC, Nickels KC. Electroclinical syndromes: infantile
onset. In: Swaiman KF, Ashwal S, Ferriero DM, Schor NF,
Finkel RS, Gropman AL, Pearl PL, Shevell MI, editors.
Swaiman’s Pediatric Neurology. Sixth Edition: Elsevier; 2017. p.
557–68.
[11] Niedermeyer E, Lopes da Silva F. Electroencephalography. 2nd
ed. Urban and Schwarzenberg;1987.
[12] Zou G. A modified poisson regression approach to prospective
studies with binary data. Am J Epidemiol 2004;159:702–6.
[13] Alvarez N, Lombroso CT, Medina C, Cantlon B. Paroxysmal
spike and wave activity in drowsiness in young children: Its
relationship to febrile convulsions. Electroencephalogr Clin Neu-
rophysiol 1983;56:406–13.
[14] Berg AT, Shinnar S, Darefsky AS, Holford TR, Shapiro ED,
Salomon ME, Crain EF, Hauser AW. Predictors of recurrent
febrile seizures. A prospective cohort study. Arch Pediatr Adolesc
Med 1997 Apr;151(4):371–8.
[15] Y}ucel O, Aka S, Yazicioglu L, Ceran O. Role of early EEG and
neuroimaging in determination of prognosis in children with
complex febrile seizure. Pediatr Int 2004;46:463–7.
[16] Karimzadeh P, Rezayi A, Togha M, Ahmadabadi F, Der-
akhshanfar H, Azargashb E, Khodaei F. The best time for
EEG recording in febrile seizure. Iran J Child Neurol 2014;8:20–5.