Epilepsy following febrile seizures (FS) has been estimated between 2% and 7%. It concerns a prospective study in a large sample of children with a long-term follow-up. The aim of this study is to identify the prognostic factors that can lead children with FS to epilepsy.
Children with a first episode of FS were included. We gathered information about prenatal and perinatal history, family history of FS and epilepsy in first- and second degree relatives, age at the time of the initial FS, dates of FS recurrences, focality, duration of the FS and recurrent episodes within the same febrile illness, height and duration of fever prior to the seizure, cause of the fever, and frequency of febrile illnesses. Patients were seen every 4–6 months and also at each recurrence.
A group of 560 children with a first FS met all entry criteria. Epilepsy was recorded at 5.4%. Statistical analysis was performed between children with epilepsy and those with no afebrile seizure. We analyzed FS recurrences in accordance with the occurrence of epilepsy. From the third FS recurrence and beyond, only focality continued to have prognostic value.
Main prognostic factors for the development of epilepsy after FS are: (1) complex FS that increased the risk for epilepsy 3.6 times, (2) age at onset of FS beyond the third year of life that raised the risk 3.8 times, (3) positive family history of epilepsy 7.3 times, and (4) multiple episodes of FS about 10 times. Focality at the first and the second FS recurrence increased the risk of epilepsy about 9.7 and 11.7 times, respectively. Focality was the only factor that continued to be significant in further FS recurrences. A prognostic profile of each child with FS would be very useful for the follow-up of these children.
Frequency of febrile seizures (FS) in general pediatric population ranges between 2% and 5% and concerns mainly children from 6 months to 5 years of age (Berg et al., 1997; AAP, 2011). Epidemiologic studies have shown that the recurrence risk of FS depends on genetic background. Gene analyses have located the responsible genes on chromosomes 2, 5, 8, 19p, and 19q (Winawer & Shinnar, 2005; Kang et al., 2006). However, the mechanisms underlying the genesis of FS are still unknown. The risk of having a first recurrence is estimated between 23% and 42%. Approximately one third to one half of those children with a first FS recurrence, will have a second recurrence, and only 10% will have three or more recurrences (Pavlidou et al., 2008). Epilepsy after FS has been investigated for many years, with conflicted results. Epilepsy following FS has been estimated between 2% and 7%, and is four to five times greater than that of the general pediatric population (Neligan et al., 2012).
However, all children with a history of FS do not have the same risk of subsequent epilepsy. This is a prospective study in a large sample of children with a long-term follow-up. The aim of this study is to identify those prognostic factors that can lead children with FS to epilepsy. This knowledge could be very useful for the management and follow-up of these children.
Patients and Methods
Children with a first episode of FS were included. The children were seen either at the outpatient pediatric neurology clinic or were hospitalized for their first episode of FS. In our area, all children with FS (simple or complex) almost are either referred to a pediatric neurology outpatient clinic or hospitalized. Epilepsy was defined as the occurrence of at least two episodes of afebrile seizures in two different days. The types of epilepsy that occurred after FS were designated according to the latest classification of epilepsies and epileptic syndromes (Berg et al., 2010).
Criteria for eligibility were the following: no previous personal history of afebrile or neonatal seizures; no evidence of intracranial infections or head trauma; no previous anticonvulsant medication; and age between 3 months and 6 years. None of the children received any antiepileptic treatment, including intermittent diazepam prophylaxis, during the study period.
The parents of the children were interviewed by a research assistant. The interviewer elicited a complete description of the seizure. Descriptions of the seizures were compiled from the medical records and the interview, and each seizure was classified on the basis of this information by an expert in pediatric neurology. According to the protocol, we gathered information about prenatal and perinatal history of each child, age at the time of the initial FS, dates of FS recurrences, presence of focal features (including Todd's paresis), duration of the FS episode and recurrent episodes within the same febrile illness, height of temperature and duration of fever prior to the seizure, cause of the fever and frequency of febrile illnesses during the follow-up period, and family history of FS and epilepsy in first- and second-degree relatives. We specifically analyzed the data from the children with a history of maternal epilepsy and children with no history of maternal epilepsy regarding the subsequent development or not of epilepsy. All of the above factors were evaluated regarding their prognostic value for subsequent epilepsy in children with FS. Body temperature (axillary) was recorded in degrees of Celsius. Temperature up to 38.5°C was defined as low, whereas greater than 38.5°C as high. At the time of the first FS a detailed neurologic examination was conducted by a pediatric neurologist. Parents were fully informed about the nature and management of FS. All children had electroencephalography (EEG) 7–10 days after the FS episode and while the child was afebrile. Lumbar puncture was performed only in cases that were indicative of meningitis or encephalitis, whereas neuroimaging (computerized tomography [CT] or mainly magnetic resonance imaging [MRI]) was performed on children with focal seizures, febrile status epilepticus, or abnormal neurologic examination.
The median time of follow up was 5.3 years (range 2.5–8.5 years, standard deviation [SD] ± 1.3 years). During the follow-up period, at every recurrence clinical reevaluation and a new EEG were performed. The patients' parents were fully informed about the nature and management of FS. Parents and patients were seen every 4–6 months, in order to reinforce the study program, and also at each recurrence they called the study staff and provided all the necessary information about the FS recurrence or the occurrence of an afebrile seizure.
A univariate analysis was performed between the examined factors at the total episodes of FS and also to each recurrence separately in correlation to the occurrence or not of subsequent epilepsy. We investigated several factors on whether they can predict subsequent epilepsy, with the Pearson chi-square test (p-value < 0.05). We identified index “γ”, which showed how powerful the connection was between the variables that we examined. Finally, we proceeded to a multivariate analysis (Cox regression-survival analysis), in order to identify the factors that contribute to the prognosis of epilepsy. The relative risk (RR) was calculated, which showed how many times a certain factor increase the risk for subsequent epilepsy. Statistical analysis was performed using SPSS 17.0 software (IBM Corp., Armonk, NY, U.S.A.).
A group of 560 children with a first episode of FS met all entry criteria. However, 59 children were excluded because no follow-up information could be obtained after the initial evaluation. This report is, therefore, based on the remaining 501 children. The age at onset of FS ranged between 3 months and 5 years old (median age at onset 23 months ± 12 months). Four hundred sixty-three (92.4%) of 501 children had normal perinatal history, 3 (0.6%) of 501 had twin pregnancy, 11 (2.2%) of 501 had prematurity (<36 weeks of gestation), and 24 (4.8%) of 501 had pathologic perinatal history (Apgar score < 7 or hospitalization at the neonatal intensive care unit for more than 3 days). None of the children revealed psychomotor retardation or regression during the follow-up period.
The cause of fever at the initial episode of FS was attributed to viral infections at 88.6% of the cases, to microbial ones at 10.4%, and after immunization at 1%. One hundred fifty-four (30.7%) of 501 children had frequent febrile infections (more than four at 6 months).
The EEG at the first episode of FS was pathologic in 105 (21%) of 501 children. From those children, only 4 had epilepsy. Generalized spike-and-wave complexes were recorded in 58 children; 32 had focal epileptiform discharges mostly from the centrotemporal and occipital regions, and 15 showed nonepileptiform abnormalities. It seems that there is a correlation between age and EEG abnormalities. From the 58 EEG studies with generalized discharges, 46 belonged to children older than 3 years of age, whereas 12 were recorded in children between 25 and 36 months of age. We did not find any specific age correlation on the EEG studies with focal discharges. At the total EEGs at FS recurrences (n = 406) only 61 (15%) had pathologic findings. From those 61 children with pathologic EEG findings, only two developed later on epilepsy.
Lumbar puncture was performed for diagnostic purposes in 45 children and all were normal. Brain CT was performed in 11/501 children and all were within normal limits, whereas brain MRI was performed in 54/501 children and pathologic findings were found in 9.
From the study analysis, 221 (44.2%) of 501 children had at least one FS recurrence. Only one recurrence was reported in 115/221 children, whereas 106/221 recurred for a second time. Three recurrences occurred in 48/106 children, and from those 31 children had four or more recurrences. During the first 6 months of the study there were 55% of the total FS recurrences, 70% in the first 12 months, 90% appeared during the first 2 years of follow-up.
Epilepsy occurred in 27 (5.4%) of 501 children. Only 14 (2.8%) of 501 children had one episode of afebrile seizures during the study period, whereas 460 (91.8%) had no afebrile seizures.
The time period of occurrence of epilepsy since the first FS ranged from 6 to 42 months, with a median of 30 ± 8 months. The majority of the cases were led to epilepsy during the first 3 years from the initial episode of FS.
Statistical analysis was performed between children with epilepsy (n = 27) and those with no afebrile seizure (n = 460). The 14 children with only one episode of afebrile seizure were excluded from the statistical analysis. Univariate analysis showed that the following were the significant prognostic markers for subsequent epilepsy: (1) positive family history of epilepsy and especially maternal preponderance, (2) complex FS in general, (3) focal FS, (4) Todd paresis, (5) short duration of fever (<12 h) prior to the first episode of FS, and (6) multiple episodes of FS (≥4) (Table 1). The remaining factors that were examined were not statistically significant.
Table 1. Prognostic factors for subsequent epilepsy (univariate analysis)
Pearson χ2 test
Two sided p-Value
Family history of epilepsy
Maternal history of epilepsy
Fever <12 h prior to the first FS
According to the multivariate analysis, the following were powerful prognostic factors for subsequent epilepsy after FS: (1) multiple episodes of FS (four or more), (2) positive family history of epilepsy and especially maternal preponderance, (3) late age at onset of FS (over 3 years of age), and (4) complex FS (focality, recurrence during the same febrile illness, duration more than 15 min) (Table 2). Complex FS increased the risk for epilepsy following FS 3.6 times, age at onset of FS beyond the third year of life 3.8 times, positive family history of epilepsy 7.3 times, and multiple episodes of FS about 10 times.
Table 2. Prognostic factors for subsequent epilepsy (multivariate analysis)
Age at onset of FS (>36 months)
Family history of epilepsy
We also analyzed FS recurrences in accordance with the occurrence of epilepsy. Our analysis reached as far as four recurrences. We did not proceed to further recurrences, since the number of the children that had more than five recurrences and developed finally epilepsy was very small. We compared, at each recurrence separately, the characteristics of the children with no afebrile seizure and those with epilepsy. The elements of FS that we studied at each recurrence were the type and duration of each seizure, age at each recurrence, hours of fever prior to FS, and recurrence of FS within the same febrile illness.
A first FS recurrence was recorded in 221 children. From those 221 children, 23 revealed epilepsy and 184 did not have any episode of afebrile seizures. One hundred six children had a second FS recurrence. From those 106 children, 19 children had later epilepsy and 76 did not have any episode of afebrile seizure. Forty-eight children that had a third FS recurrence. From those 48 children, 11 were diagnosed with epilepsy and 26 did not have any episode of afebrile seizure. Four or more recurrences of FS occurred in 31 children. From those 31 children, 8 had subsequent epilepsy, whereas 18 did not show afebrile seizures.
Multivariate analysis of the first FS recurrence revealed that the prognostic factors for subsequent epilepsy are focality and short duration of fever prior to the occurrence of the seizure. More specifically, the presence of focal seizures at the first FS recurrence increased the risk for subsequent epilepsy by 9.7 times, whereas short duration of fever prior to the first FS recurrence increased the risk of epilepsy about 2.5 times. Those children that at first FS recurrence had both focality and short duration of fever had an increased risk for epilepsy of about 8.5 times (Table 3). The remainder of the factors did not influence the risk for subsequent epilepsy.
Table 3. Prognostic factors of subsequent epilepsy at FS recurrences
>15 min duration
Age at onset >3 years
<12 h of fever
Recurrence at the same febrile illness
Focal seizures + short duration of fever
Data for the second FS recurrence revealed as significant prognostic factors again focal seizures and short duration of fever. As shown in Table 3, the presence of focal seizures at the second FS recurrence increased the subsequent risk of epilepsy 11.7 times, whereas short duration of fever prior to the second FS recurrence increased the risk about 1.8 times. Children with both focality and short duration of fever before FS recurrence had an increased risk of subsequent epilepsy of about 4.9 times. The remainder of factors was not found to be statistically significant.
Multivariate analysis at the third FS recurrence showed that only focality carried prognostic value regarding the development of epilepsy. The rest of the factors were not statistical significant. Focality was the only factor that continued to be significant even at four FS recurrences, increasing the risk of epilepsy up to 30 times. The types of epilepsy that occurred in 27 children of the study were generalized tonic–clonic epilepsy in 11 children, rolandic epilepsy in 5, other focal epilepsies in 3, generalized epilepsy with FS plus (GEFS+) in 4, petit mal in 3, and juvenile myoclonic epilepsy in one child. Of the 19 children that finally developed generalized epilepsy, 15 had a history of only simple febrile seizures. Of the eight children who finally developed focal epilepsy, six children had experienced focal FS (p-value 0.024).
Epilepsy is the main serious complication of FS. Although this issue has troubled many authors through the past years, the studies that exist are few and contradictory. Most of these studies are retrospective and concern mainly teenagers or adults with epilepsy and a history of FS (Tsai & Hung, 1995; Leung & Robson, 2007).
It is common knowledge that manifestation of epilepsy is more frequent in children with FS versus the general pediatric population. The risk of epilepsy after FS depends mainly on the type of FS and the duration of follow-up (Annegers et al., 1987; Berg & Shinnar, 1996). Most studies report a common pathogenetic mechanism or a cause–effect relationship between FS and epilepsy (Verity et al., 1993; MacDonald et al., 1999; Bender & Baram, 2007). Recent clinical and molecular gene studies reinforce the opinion that FS and epilepsy share a powerful genetic connection (Mancardi et al., 2006; Abou-Khalil et al., 2007; Dube et al., 2009). Gene analyses in families with epilepsy identified mutations at genes that are responsible for ion channels and recognized a large range of phenotypes, from FS to severe forms of childhood epilepsy (Burgess, 2005). Vestergaard et al. (2007) reported that siblings of children with FS frequently display epilepsy, even if they did not experience FS. This indicates that genes and environmental factors common to the members of the same family, when they are expressed differently, contribute to the occurrence of FS and epilepsy. There are also indications that FS of long duration can probably harm the developing brain and provoke epilepsy (Berkovich & Scheffer, 1998; Saltik et al., 2003).
There has been much interest in preventing FS and reducing their recurrence risk by intermittent treatment with benzodiazepines (Rosman et al., 1993; Verrotti et al., 2004; Khosroshahi et al., 2011). On the other hand, there are conflicting results from other studies, which show no significant seizure control with the above treatment (Autret et al., 1990; Uhari et al., 1995; Capovilla et al., 2009). Other studies have shown that only children that are at high risk for recurrent multiple FS benefit from use of intermittent diazepam ,that is, giving diazepam during fever (Pavlidou et al., 2006). Recent meta-analysis found no clinically important benefits for children with FS by intermittent prophylaxis (Offringa & Newton, 2012). Multiple episodes of FS could be a reflection of a lower seizure threshold (Knudsen et al., 1996).
Recent studies report that epileptiform discharges on the EEG studies of patients with FS are important predictive markers for the development of epilepsy (Kanemura et al., 2012; Kim et al., 2013; Wo et al., 2013). The recent FEBSTAT study showed that focal slowing or attenuation are present in EEG recordings obtained within 72 h of the febrile status epilepticus in a substantial proportion of children and are highly associated with MRI evidence of acute hippocampal injury (Nordli et al., 2012).
According to our study, the main prognostic factors for subsequent epilepsy after FS are
Multiple FS episodes
In the present study, multiple episodes of FS (four or more) were the most powerful factor for the prognosis of subsequent epilepsy, especially in multivariate analysis. This element is confirmed also by other authors, who report that the incidence of epilepsy increases as the number of FS increases (Berg & Shinnar, 1996; Vestergaard et al., 2007).
In a retrospective study of Annegers et al. (1987), the number of FS was an important prognostic factor in the univariate analysis, but when it was examined in relationship with other factors its prognostic value was reduced. It is not totally clear why multiple episodes of FS are related to the increased occurrence of epilepsy. A possible explanation is that the number of FS could probably be an indicator of sensitivity toward the genesis of seizures. This opinion is even more reinforced from other studies that support that every new episode of seizures disturbs the “electrical” balance and triggers a new seizure (Berg & Shinnar, 1997; Vestergaard et al., 2007). Another possible explanation is bias selection, since most of the studies analyzed data from tertiary hospitals with more severe episodes of FS (i.e., focal or prolonged). Our study did not have a problem of bias selection, since in our region all children with FS, and not only the severe cases, are either hospitalized or attended at the pediatric neurology outpatient clinic.
Positive family history of epilepsy
Positive family history of epilepsy of first- and second-degree relatives was found to be a major prognostic factor. Literature is plain for the positive predictive value of family history of epilepsy (Verity & Golding, 1991; Sadleir & Scheffer, 2007). In addition, our multivariate analysis showed that maternal history of epilepsy is a strong prognostic factor. To the best of our knowledge, this finding is described for the first time in literature. We consider, that this finding is of major significance, especially in regard to the genetics of epilepsy in children with FS.
Complex FS, not only at the first episode but also at recurrences, plays an important role in the prognosis of epilepsy. Other studies have shown that all the characteristics of complex FS (focality, duration >15 min and recurrence within the same febrile illness) are prognostic markers for the development of afebrile seizures (Sfaihi et al., 2012; Wo et al., 2013). Few studies have analyzed separately the prognostic value of each of the three features of complex FS. Annegers et al. (1987) reported that the risk of subsequent epilepsy in children with one of the three characteristics of complex FS is about 6–8%, with two 17–22%, and with three up to 49%.
Late age at onset of FS
Our study showed that late age at onset of FS (>3 years old), as an isolated factor, has no influence at the occurrence of epilepsy, but only when it is taken into account with other prognostic markers, plays an important role in epileptogenesis. Specifically, late age at onset of FS increased the risk of subsequent epilepsy about four times. A retrospective study found that children who had their first FS at an age <1 year and >3 years had epilepsy more often in the first 2 years after the FS episode (Vestergaard et al., 2007). Other authors found that children with a first FS at an age younger than 12 months old had a higher incidence of epilepsy (Nelson & Ellenberg, 1976; Annegers et al., 1987).The interpretation of this finding is that severe episodes of FS tend to occur at a younger age.
The present study showed that focality, not only at the first FS but mostly at recurrences, is a reliable prognostic factor for subsequent epilepsy. Indeed, the prognostic power of focality was increased at each recurrence. Therefore, children with focal seizures at the fourth recurrence (i.e., more close timely to the occurrence of epilepsy) had an increased risk of epilepsy of up to 30 times. Other authors also supported that the most powerful prognostic factor between the three characteristics of complex FS for subsequent epilepsy was focality (Verity & Golding, 1991).
Our results have shown that FS recurrence within the same febrile illness cannot predict subsequent epilepsy. In addition, the number of febrile status epilepticus in our study was restricted and children with so we could not reach to statistical significant results. Sapir et al. (2000), reported that none of their patients with prolonged FS developed later epilepsy. The same conclusion is reached also by other authors, as they noted that febrile status epilepticus does not increase the risk either of FS or afebrile seizures (Maytal & Shinnar, 1990). A recent study concluded that prolonged FS is the result of a previous insult of the hippocampus and not the cause of it (Scott et al., 2006).
Different results are published by Verity et al. (1993), who reported that the incidence of subsequent epilepsy was higher in children with febrile status epilepticus. Saltik et al. (2003) found that 13.7% of their patients with epilepsy had a history of febrile status epilepticus. However, most of the studies differ from ours, since they concern patients hospitalized at reference centers and therefore involved a select group of severe cases, who possibly had a preexisting brain injury. It has been proven that prolonged seizures can harm the neurons. The question that arises is how often such an injury can occur and why it happens to certain children. It seems that the neuronal insult is connected with the cause of the seizures and not with the seizures per se (Verity, 1998).
Most of the studies in the literature have examined various prognostic factors for subsequent epilepsy, either the total of FS episodes (i.e., first FS and recurrences) or only the first episode of FS. We analyzed each recurrence, since we hypothesized that recurrences could change the prognostic profile for the child regarding the development of epilepsy. Our findings suggest that focality and short duration of fever (<12 h) prior to the seizures have a significant prognostic value for the first and second recurrence, whereas their combination increase much more the risk of epilepsy. By the third recurrence, the only factor that remained statistically significant was focality, which increased its statistic power with further recurrences.
The types of epilepsy that follow FS are usually generalized tonic–clonic epilepsy, temporal lobe epilepsy, and less often juvenile myoclonic epilepsy and benign epilepsy with centrotemporal spikes. Several studies have shown that there is no difference in the incidence of generalized and focal epilepsy between the children with and without a history of FS (Annegers et al., 1987; Verity & Golding, 1991; Wo et al., 2013).
These findings support the hypothesis that FS do not contribute more to the occurrence of focal epilepsy, except of the group of hippocampal sclerosis and temporal lobe epilepsy. Temporal lobe epilepsy is more associated with long duration of FS and less with focality (Heuser et al., 2011). The relationship between FS and mesial temporal sclerosis has been one of the most contentious issues in epileptology. A few authors suggest that prolonged FS directly damage the hippocampus and surrounding structures, and others that damage predates the FS and causes seizures, both with fever and afebrile seizures (Neville & Gindner, 2010; McClelland et al., 2011).
Another noteworthy syndrome with FS is generalized epilepsy with FS plus (GEFS+). It is inherited by an autosomal dominant trait and is characterized by a heterogeneous phenotype that includes FS continuing beyond the sixth year of life and/ or intermingled with afebrile seizures. Temporal lobe epilepsy is rarely a phenotype of GEFS+ (Abou-Khalil, 2010).
The tendency of few children to have focal FS, probably reflects an already existing neurologic insult, which is also responsible for the occurrence of subsequent focal epilepsy (Rocca et al., 1987). In our study, most of the children who manifested generalized epilepsy had previous simple FS, whereas the majority of children who revealed focal epilepsy had a history of complex FS. We followed the children of our study for a median of 5.3 years. However, time plays an important role in epileptogenesis after FS and it seems that the risk of epilepsy continues to increase through the years (Annegers et al., 1987). Therefore, we should acknowledge that epilepsy can occur after many years from the initial episode of FS.
The mechanisms by which FS can contribute to the development of epilepsy are not fully resolved. Among the key remaining goals is the development of early and effective biomarkers that will define which child with FS is at risk for epilepsy.
The present study intends to determine a prognostic profile in the child with FS regarding the risk for subsequent epilepsy. Maternal preponderance is for the first time described in the literature as an important risk factor for subsequent epilepsy after FS. In addition, the timing and the nature of FS recurrences have impact toward the direction of epileptogenesis.
A challenge for the clinician is the early recognition of those children who are at increased risk for developing epilepsy after FS. To the best of our knowledge this is the first prospective study about the prognosis of epilepsy in children with FS that analyzes the data not only of the first episode of FS, but also of each recurrence separately.
We thank all the children of the study and their families for their patience and support.
None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with the guidelines of the Committee on Publication Ethics (COPE) guidelines for ethical publication (http://publicationethics.org/).
Dr Efterpi Pavlidou is a Consultant Pediatrician working in the field of Pediatric Neurology at a tertiary hospital since 2011. She has finished her dissertation in pediatric neurology in 2007.