Prolonged febrile seizures, clinical characteristics, and acute management
Child Neurology and Development Unit, Tel-Aviv Sourasky Medical Center, Tel-Aviv, Israel
Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel
Address correspondence to Haim Bassan, Child Neurology and Development Unit, Dana Children's Hospital, Tel-Aviv Sourasky Medical Center, 6 Weizmann Street, Tel-Aviv 64239, Israel. E-mail: email@example.com
Prolonged febrile seizures (PFS) lasting ≥15 min have been associated with increased risk for epilepsy in later life. Initial treatment, mostly prehospital, aims to prevent its evolution to febrile status epilepticus (FSE) and reduce adverse outcome. Paucity of information is available on the immediate treatment before reaching a hospital facility.
We obtained data, prospectively, on all children who presented from January 2008 to March 2010 with PFS to the emergency rooms of four Israeli medical centers. Information related to seizure semiology, treatment, and medical history was collected into a predefined pro forma form and reviewed centrally.
Sixty children, median age 18.3 months (interquartile range [IQR] 12–28) were included with a median seizure duration of 35 min (IQR 26–60), 43 (71.7%) lasting ≥30 min. Seizures had focal onset in 34 infants (57%). Fifty-four families (90%) activated the ambulance service; median ambulance arrival time was 8 min (IQR 5–10), 33 (61%) were medically treated by the ambulance paramedic, of whom 15 (45%) responded to treatment. Twelve children with active seizures did not receive medications. Initial treatment with rectal diazepam was more common in those with seizure duration >30 min.
Most children with PFS are treated with antiepileptic drugs early by the ambulance service. However, even timely treatment does not prevent status epilepticus in the majority of cases. These data highlight the need for effective early treatment of this common pediatric emergency.
Prolonged febrile seizures (PFS) is an important subtype of complex febrile seizures. Together with “focality” and “multiple seizures” subtypes, complex febrile seizures account for approximately 25–30% of all febrile seizures (Berg & Shinnar, 1996; Waruiru & Appleton, 2004; Hesdorffer et al., 2011). PFS have been associated with increased risk of epilepsy, recurrent febrile seizures, and febrile status epilepticus (FSE). There has been some debate whether 10 or 15 min should be the cutoff epidemiologically with more recent data, suggesting that 10 min is a more appropriate cutoff (Hesdorffer et al., 2011). But in terms of screening for those at risk for status epilepticus and looking at whether treatment is effective, 15 min remains a reasonable definition (Shinnar et al., 2008).
Status epilepticus (SE) is defined as a seizure lasting beyond 30 min or a series of seizures lasting a total of 30 min or more, without full recovery to baseline (Commission on Epidemiology & Prognosis, ILAE, 1993; Dodson et al., 1993; Riviello et al., 2006). Febrile status epilepticus (FSE) meets the definition of both a febrile seizure and status epilepticus and accounts for 5–9% of children experiencing febrile seizures (Berg & Shinnar, 1996; Shinnar et al., 2008; Hesdorffer et al., 2011). FSE is the most common type of status epilepticus in children, accounting for two thirds of all cases with status epilepticus in the second year of life and one fourth of all childhood status epilepticus (Shinnar et al., 1997; Raspall-Chaure et al., 2007). FSE is a medical emergency as are all forms of status. With FSE in particular, the longer the FSE continues, the less likely it is to spontaneously stop (Shinnar et al., 2008), posing additional risks of morbidity and mortality. A recent report in a large U.S. cohort (the FEBSTAT study) suggested that FSE is often not well recognized (Shinnar et al., 2008; Hesdorffer et al., 2012).
There is a growing body of evidence on the association between prolonged febrile convulsions, and hippocampal injury, subsequent mesial temporal sclerosis, and temporal lobe epilepsy (VanLandingham et al., 1998; Lewis et al., 2002; Shinnar, 2003; Shinnar et al., 2012). In light of these concerns, the acute management of prolonged febrile seizures becomes highly important, particularly the initial management at home, in the ambulance, or in the emergency room. We report on a prospective study of the acute management of PFS in Israel.
In this prospective multicenter study, we collected data from January 2008 to March 2010 on all consecutive children who presented to the emergency departments of four Israeli hospitals with a febrile seizure lasting more than 15 min. Febrile seizure was defined as a provoked seizure where the sole acute provocation was fever (temperature 38.4°C, 101.0°F), no history of afebrile seizures, and no evidence of an acute central nervous system (CNS) infection or insult (National-Institute-of-Health, 1980; Commission on Epidemiology & Prognosis, ILAE, 1993).
Data obtained included the following: (1) demographic data, past medical history including perinatal and developmental details, as well as family history of epilepsy; (2) pre-hospital management information; and (3) information from the emergency room including timing of arrival to hospital, and overall duration of seizures, defined as the interval between the reported onset of seizure and the cessation of clinical seizure activity. In case of multiple but discrete seizures with full recovery in between, the duration of the longest seizure was used in the analysis. In addition, we determined the seizure semiology, etiology of fever, and type of medication used.
At each center, the research team collected copies of the ambulance transport service notes, and the emergency department nursing and doctors' datasheets. In cases where the children were hospitalized, records from the pediatric department or intensive care unit were also obtained. The seizure narrative was recorded and the semiology of the seizure(s) whether focal or generalized was determined by each center clearly on clinical grounds, and then confirmed by two of the investigators (M.B. and H.B.) who reviewed all cases and reached consensus. To ensure consistency with other studies, one investigator (S.S.) who has been involved with central scoring of several studies (Shinnar et al., 2008) also reviewed a subset of cases and conferenced with the central reviewers.
Seizures were categorized according to the International League Against Epilepsy (ILAE) classification (Commission on Epidemiology & Prognosis, ILAE, 1981; Berg et al., 2010) as focal onset or generalized using the same approach and criteria as in the FEBSTAT study (Shinnar et al., 2008; Hesdorffer et al., 2012). Seizures were further characterized as continuous or intermittent.
The development of each child was graded as normal or delayed for each of the motor, language, and social domains, based on a structured parental interview regarding developmental milestones, need for services or special education, and a neurodevelopmental evaluation performed by an experienced child neurologist. No formal developmental testing was administered. A significant neurodevelopmental disorder was defined as cerebral palsy and/or a significant developmental disability of at least two developmental domains. In all cases, the attending neurologists who were involved in the care of the seizing child filled out the study forms.
The study was approved by the institutional review boards at all participating institutions.
We categorized our data according to the semiology of seizure(s), whether focal or generalized, and additionally according to the duration of seizures (<30 min vs. ≥30 min). We chose the cutoff of 30 min because it represents the threshold for diagnosis of FSE. Comparisons for continuous variables were made by Student's two-tailed t-test and for dichotomous variables by Fisher's exact test. To examine the independent effect of several explanatory variables, we constructed a multivariate logistic regression model that predicts the probability of having a prolonged (≥30 min) seizure. SPSS (SPSS, Inc, Chicago, IL, U.S.A.) was used for all computations.
Study population and seizure semiology
Sixty children with a median age of 18.5 months (interquartile range [IQR] = 12–28 months) met the study inclusion criteria. Demographic and medical data are presented in Table 1. Six (10%) had a history of perinatal complications; 15 (25%) had a prior febrile seizure, of whom 5 were prolonged. Eleven (18.3%) had a significant neurodevelopmental disorder. Median duration of seizures was 35 min (IQR = 26–60 min).
Table 1. Study population, demographics, and medical history
Clinical characteristic (n = 60)
Number (percent); Median (IQR)
Significant neurodevelopmental disorder was defined as cerebral palsy and/or developmental delay of at least two domains (see 'Methods' section).
Thirty-four children (57%) had a focal onset seizure, of whom 22 had a definite focal onset and 12 had probable focal onset (Table 2). In addition, there were four children whose seizures were clinically classified as generalized onset but who had a focal finding on their electroencephalography study (Table 2). Compared to generalized PFS (n = 26), focal PFS (n = 34) were associated with lower gestational age (mean ± standard deviation [SD], 37.3 ± 4 vs. 39.4 ± 3 weeks, p = 0.03), lower birth weight (mean ± SD, 2727.3 ± 781 vs. 3164 ± 553.4 g, p = 0.01) and a significant neurodevelopmental disorder (10 [29.4%] vs. 1 [3.9%], p = 0.01).
Of the 60 children with PFS, 12 (20%) with a median seizure duration of 28 min (IQR 19.5–35) spontaneously stopped during the arrival of the ambulance. Three children were given rectal diazepam by the parents and four received early treatment at a local urgent care station before transport. Of the 60 children, 54 (90%) were brought in by ambulance. Of the 41 actively having seizures in the ambulance, 33 were treated in the ambulance, whereas 8 were not treated because seizures were not recognized. One of these eight children received rectal diazepam at home prior to arrival of the ambulance. The four actively convulsing children who were transported by family car or taxi were also not treated prior to arrival to the emergency department. The prehospital treatment is summarized in Figure 1 and Table 3.
Table 3. Prehospital management of prolonged febrile seizures
Diaz (rectal) followed by Diaz or Midaz (intravenous)
Diaz or Midaz (intravenous)
Midaz (intramuscular or intranasal)
Seizures stopped by ambulance treatment
Seizures upon arrival to ER
For those children who were treated in the ambulance, the antiepileptic drugs (AEDs) given included intravenous diazepam or midazolam (n = 19), intranasal or intramuscular midazolam (n = 5), and rectal diazepam solely or followed by intravenous benzodiazepine (n = 9). Lorazepam is less available in Israel and was not used by the ambulance service (Table 3). In 24 children, the doses of AEDs were available. For intravenous midazolam the median dose was 0.16 mg/kg/dose with a range from 0.03 to 0.23 mg/kg/dose, whereas for rectal diazepam the median dose was 0.5 mg/kg/dose with a range of 0.42–0.5 mg/kg/dose. None of the children had respiratory depression. AED therapy was successful in terminating the seizure prior to arrival at the emergency department in 15 cases (45.5%). Those receiving rectal diazepam were less likely to respond. Only one (11%) of nine children receiving rectal diazepam (in that case followed by intravenous benzodiazepine) responded prior to arrival in the emergency department compared with 11 (58%) of 19 children who received intravenous diazepam or midazolam (p = 0.02). However, the group initially treated with rectal diazepam was younger (median age 14.5 months) than the group of 19 children initially treated with intravenous benzodiazepines (median age 25 months) (p = 0.06).
Thirty-one children (52%) were still seizing upon arrival to the emergency department (ED). Thirty received immediate AED treatment and in one child (with intermittent seizure) seizures were not appreciated by the ED staff and finally stopped spontaneously. The ED doctors used multiple AED treatments (two to four drugs) in 15 children, and 38 children were admitted to the hospital. EEG reports were available in 37 children (61.7%) (20 normal, 17 abnormal). None had bedside continuous EEG. A referral to an EEG was not associated to the child age, seizure duration, focality, and to multiple seizures subtype (p = NS). Lumbar puncture (LP) was performed in 12 patients (20%), ages 3–35 months; half of them were younger than 18 months of age and none showed evidence of bacterial meningitis. In addition of having a younger mean age (months) (17.2 ± 9.1 vs. 27.3 ± 25.2, p = 0.03), children who underwent an LP tended to have a longer mean duration of seizure (min) (73.8 ± 36.6 vs. 42.4 ± 27, p = 0.001), and therefore most of them had active seizures upon arrival to ED (11/12 [91.7%] vs. 21/48 [43.75%] p = 0.003). Focality and multiple seizures subtypes were not associated with a performance of LP (7/12 [58.3%] vs. 27/48 [56.3%], p = 0.6; 3/12 [25%] vs. 6/48 [12.5%], p = 0.25, respectively).
Predictors of prolonged febrile seizures
We compared the characteristics of seizures <30 min versus seizures 30 min or longer and found that the latter were more common in young children (20.7 ± 14 vs. 36.9 ± 36 months, p = 0.01) and associated with rectal diazepam treatment solely or followed by intravenous benzodiazepines (p = 0.02), Table 4. We then constructed a multivariate logistic regression model incorporating risk factors that reached a p-value < 0.2 on the univariate analysis. We included in the model: age, history of neonatal seizures, intermittent seizure semiology, and type of AED. This model revealed an independent effect of intermittent semiology (p = 0.02) and initial rectal diazepam treatment (p = 0.001) on prolonged (≥30 min) seizure activity.
Table 4. Prolonged febrile seizures, below or above the 30 min threshold
Duration of seizure
<30 min (n = 17)
≥30 min (n = 43)
Diaz, diazepam; Midaz, midazolam.
Seizure duration, min
21.1 ± 4.6
59.6 ± 30.8
36.9 ± 36
20.7 ± 14
Gestational age, weeks
37.8 ± 3.6
38.4 ± 3.6
Birth weight, g
History of neonatal seizures
Significant neurodevelopmental disorder
Previous febrile seizure
Multiple with interictal recovery
Intermittent without interictal recovery
Time until ambulance arrival
8.6 ± 5.2 (n = 14)
9.6 ± 7.2
Drug treatment in ambulance
Diaz (rectal) followed by Diaz or Midaz (intravenous) vs.
Diaz or Midaz (intravenous) or Midaz (intranasal, intramuscular)
We found that over 70% of children with PFS >15 min end up with FSE defined as >30 min even with treatment. This is consistent with recent data that there are two populations of febrile as well as afebrile seizures: short seizures that are likely to resolve before 10 min (Shinnar et al., 2001a2001b; Hesdorffer et al., 2011) and those with FS >10 min are likely to be prolonged.
The median PFS duration in our study is 35 min shorter than that reported in the FEBSTAT study (median 68.0, IQR 47.0–105.0) (Shinnar et al., 2008; Hesdorffer et al., 2012). This is partly due to the inclusion of children with seizures of 15–30 min duration. However, in the FEBSTAT study, all children with seizures lasting >15 min were screened (Hesdorffer et al., 2012) and there were very few who then did not meet criteria of >30 min (S. Shinnar, personal communication). Some of the differences are likely due to a faster arrival of ambulance and a much higher proportion of children receiving early aggressive treatment in the ambulance (Hesdorffer et al., 2012). Consistent with this, the median seizure duration of those who received prehospital treatment was shorter than those who did not.
Ambulance was activated by most (90%) of our population, much higher than that reported in a study in Tokyo (Sakai & Marui, 2008). It should be noted that our median ambulance response time is <10 min, which is far shorter than those reported in other studies from the United States (Jordan, 1994; Shinnar, 2007). This may reflect the high availability of ambulance service in our community and the relative proximity of residential neighborhoods to large hospitals. Unfortunately, this timely arrival of ambulance may not be applicable to other settings where ambulance service is less available or constrained by large distances or severe city traffic.
Even with a trained ambulance system that routinely treated the children, a significant number of cases were either not treated or not recognized. This too is consistent with the results of the FEBSTAT study (Shinnar et al., 2008; Hesdorffer et al., 2012). Furthermore, variation in the dosing of midazolam used by the paramedics also requires attention. Clearly, if prehospital treatment is to become the norm in children as it is in adults (Alldredge et al., 2001), more education is needed as neurologists are rarely at the front lines for treatment.
A few families approached a local ambulatory clinic for early emergency treatment before ambulance transport. Only 5% of children were treated at home by their parents with rectal diazepam. This is likely due to the fact that only five children in our population had a prior PFS, and were therefore candidates for abortive therapy.
Consistent with other studies (Berg et al., 1992; Berg & Shinnar, 1996; Shinnar et al., 2008; Hesdorffer et al., 2012), the majority of our cases were focal. The proportion of those with focal seizure is likely underestimated as four cases with clinically generalized seizures had focal EEG abnormalities. Focality was associated with neurodevelopmental impairment, suggesting that this group may carry an underlying brain dysfunction.
In agreement with other studies that showed an association between PFS and developmental delays (Shinnar et al., 2001ab; Hesdorffer et al., 2011), we found that 18% of our population had a neurodevelopmental disability. The etiology of fever was predominantly viral infections as reported in other FSE cohorts (Shinnar et al., 2008) and in febrile seizures in general.
Younger age was associated with a more prolonged seizure, and this is consistent with data that median age of children with FSE is lower than that of children with simple FS (Berg & Shinnar, 1996; Hesdorffer et al., 2012) and with animal data that the immature brain may be more susceptible to seizures (Rakhade & Jensen, 2009).
Of interest, those initially treated with rectal diazepam were more likely to have prolonged seizures than when initial administration of benzodiazepines was intravenous, intramuscular, or intranasal. We have no clear-cut explanation for the lower efficacy of diazepam when given rectally, since rectal administration of diazepam is considered to have a good bioavailability compared to intravenous administration (Cloyd et al., 1998) and a favorable efficacy (Kriel et al., 1991; Alldredge et al., 1995), and the doses used in our cohort were similar to those in the clinical trials of rectal diazepam. This result must be interpreted with caution as although prospective, this was not a randomized clinical trial and there may have been other reasons why some children received rectal and some received intravenous therapy as initial therapy. For example, as a group those receiving rectal diazepam were significantly younger, which is associated with longer seizure duration. In a nonrandomized observational study other differences may also exist, which makes comparison difficult.
However, the findings are consistent with the observations of Chin and colleagues who found that treatment of status epilepticus with intravenous lorazepam was associated with 3.7 times greater likelihood of seizure termination than treatment with rectal diazepam (Chin et al., 2008).
Lumbar puncture was performed in 20% of the patients and none showed evidence of bacterial meningitis similar to the FEBSTAT (Frank et al., 2012) and to the report by Kimia et al. (2010). In their study, of 67 children with PFS presented as a single manifestation, none had bacterial meningitis; however, of the 51 children with a focal PFS, 2 (3.9%) had bacterial meningitis. Other studies reporting on up to 17% of bacterial meningitis in children with FSE (Chin et al., 2005) and an increasing likelihood of bacterial meningitis with longer seizure duration (Offringa et al., 1992).
We acknowledge several limitations of our study. Time of onset may be underestimated in several cases, since some of the children were already seizing when witnessed by their caregivers. In addition, all children did not have EEGs during seizures, thus the timing of seizure termination was based solely on clinical judgment and was hard to define in some cases when the child enters the postictal stage. Nevertheless, the methodology used was similar to that in the FEBSTAT study and studies of prolonged seizures in the community as distinct from inpatient have all used clinical cessation of seizures as the primary end point.
Prolonged febrile seizures are not innocent events and pose a treatment challenge in the community during transport and in the emergency department. There needs to be better recognition of seizures and awareness of the need to treat them (Wheless, 2004). Randomized clinical trials similar to the PHTSE study in adults (Alldredge et al., 2001; Lowenstein et al., 2001) are needed to define the optimal therapy for the prehospital treatment in children.
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/).