Simple febrile seizures: new cut off for the duration of the crises

Our study aimed to identify a new cut-off for febrile seizure (FS) with a good prognosis, thereby replacing the 15 min described in the standard definition of simple febrile seizure (SFS). Our study was a retrospective observational study (from January 2018 to December 2018) on children admitted to the Pediatric emergency room of the Santobono-Pausilipon Hospital, Naples, Italy, Pediatric Unit of Latina, Rome, Italy, and Policlinico-Vittorio-Emanuele University Hospital, Catania, Italy, for fever, which developed SFS during the hospitalization. All included patients had their seizures classified as SFS according to the international criteria for epilepsy. We assumed a duration cut-off, and we analyzed the EEG results, neurological follow-up at 12 months, and the recurrence of the febrile seizures the following year. Then, with another calculation, we identify an optimal cut-off of 6 min. Finally, we divided the population into two groups: children with seizures having a duration greater than or less than 6 min. We found that the population with FS with a duration greater than 6 min presented EEG alteration at follow-up visits, neurological disorders, and a recurrence of FS during the following year. We suggest to introduce a new cut-off for the duration of FS that better represents the benign nature of a simple febrile event.


Introduction
Febrile seizures (FSs) are the most common neurologic disorders in childhood. FSs affect about 2-5% of all children, they mainly occur between 6 months and 5 years of age, and have a peak incidence at 18 months of age [1][2][3][4].
FSs can be classified as simple febrile seizures (SFSs), complex febrile seizures (CFSs), and febrile status epilepticus (FSE) [5][6][7]. SFSs have a duration of less than 15 min, are self-limiting, usually presenting as tonic-clonic generalized seizures, with no positive history for other neurologic disorders, with one-time occurrence in 24 h. CFSs are complex seizures lasting more than 15 min, while FSEs have a duration of more than 30 min.
While most patients with FSs should never progress to epilepsy, literature data report a slightly increased risk of subsequent epilepsy in patients with FSs: 1-2.4% of patients with SFS and 4-6% of patients with CFS might develop epilepsy later in life [4].
Factors known to be associated with a higher risk of epilepsy include: familial and personal past history of epilepsy, recurrence and duration of FSs, frontal epileptiform discharges at EEG, abnormal neuroimaging, and presence of genetic abnormalities [8,9].
Herein, the authors present a retrospective observational study on children with SFS, diagnosed according to the International League Against Epilepsy criteria [5][6][7]. We analyzed SFS lasting more than 5 min and their EEG findings. We also studied the neurological exam of children with FSs and the mean recurrence of FSs per month. Our aim was to establish the best cut-off of FS duration in order to identify "benign FSs", and differentiating FSs with better prognosis from those tending to evolve into epilepsy later in life.

Study design
A 12-month (from January 2018 to December 2018) retrospective study was performed. We randomly included patients admitted to the emergency room for fever, who developed SFSs during the hospitalization. The two centers participating to the study were: the Pediatric Emergency Unit of Santobono-Pausilipon Hospital, Naples, Italy, Pediatric Unit of Latina, Rome, Italy, and The Pediatric Emergency Unit of Policlinico-Vittorio Emanuele University Hospital, University of Catania, Italy.
All children between 6 months and 5 years of age, who were diagnosed with SFS according to the International League Against Epilepsy criteria [5][6][7] were included in this retrospective study.
Children with CFSs, a personal history of afebrile seizures, on therapy with antiepileptic medications, with a personal history of cerebral disorders and/or neurologic diseases (brain malformations, cerebral palsy, stroke, intracranial tumors, central nervous system infections or concomitant head injuries) were excluded from this study. We considered and evaluated only the seizures that occurred during hospitalization, in order to be able to correctly detect the timing of seizures. We included patients who performed an inter-critical electroencephalography (EEG) in wakefulness and in sleep regardless our will according to guidelines [5][6][7]10]. The patients' parents the showed us the result of EEG at follow-up.
For each patient, the following data were collected: age at admission, gender, personal history of FSs, familial history of epilepsy, neurological disorders; physical examination upon arrival (including body temperature, SpO2, heart, and respiratory rate); neurological examination; age at seizures onset; number of episodes before admission; associated neurological disorders; seizure semiology (according to the classification of Trinka et al. [11]); duration (the interval time between the onset and the stopping of clinical seizure activity during hospitalization monitoring by physician, excluding long post-critical seizures); EEG results at 6-month and 12-month followup; anticonvulsant therapy (nasal midazolam 0.2 mg/kg or midazolam ev 0.1-0.2 mg/kg) administered as rescue medication; number of FSs during the follow-up period; and recurrence of FS during the following year; basic antiepileptic treatment.
When EEG results had a normal background activity and absence of any electrical abnormalities, it was considered "normal". On the contrary, any pathological activity was described in detail. All the EEG studies were read by two independent neuropediatricians, and their interpretations were performed by the same two specialists, according to the International Guides on Epilepsy [5][6][7]10]. All patients' data were recorded in an Excel database and then reported into a central database to perform statistical analysis.
The ethics committee of our institution approved the study (Policlinico-OVE Ct 12,578). All research was performed in accordance with international relevant guidelines/ regulations. Informed consent was obtained from all the parents' of the studied patients. The research was performed in accordance with the Declaration of Helsinki guidelines [12].
A written informed consent was asked to the patients' parents to participate to the present study.

Statistical analysis
For statistical analysis, we used the software STATA, Stata-Corp LLC, version 14.2.
Univariate and multivariate analyses were performed considering as clinical outcomes the following variables: EEG alterations at 6 and 12 months, neurological signs at 12 months, and the recurrence of FSs one year after the inclusion in our study. Some features of the cohort were selected as covariates of the statistic model: gender, age of onset, psychomotor development, temperature, family history, and duration of seizures. These data were analyzed according to logistic regression.
Subsequently, by ROC analysis, the same outcome variables were assessed with respect to the duration of the crisis to establish an optimal cut-off value which could better determine the benignity of the disease. There are numerous methods to evaluate the optimal cut-off, but herein we have chosen to use the Youden index, which identifies the value of the test that maximizes the difference between true positive and false positive results. Once the optimal cut-off had been chosen, the study population was divided into two groups and the different clinical outcomes were analyzed by an exact Fisher test, obtaining a p-value.

Results
In our study, we evaluated 263 patients with FSs of simple nature, and 48 have been excluded according to exclusions criteria. In details, 34 pt had a history of afebrile seizures and 14 pt had on therapy with antiepileptic medications for neurological disorders (7 pt cerebral palsy, 1 stroke, 1pt intracranial tumors and 5 pt central nervous system infections). 215 children with SFSs were included in the study, 155 of the centers of Naples, 35 of the University of Catania and 25 of Latina, 115 male (53%) and 100 female (47%), with an average age of 2.2 ± 1.4 years (mean ± SD). FSs ends within minutes, typically less than 5 min. In these patients, it was not necessary to administer any drug; the seizures have been self-limiting.
Univariate and multivariate analyses considering the following variables: EEG results at 6 and 12 months, the neurological objectivity at 12 months, and the recurrence of seizure as clinical outcome, we found that the statistically most significant prognostic factor to determine whether a SFS was benign or could evolve into epilepsy was the duration of seizures.
A ROC analysis was carried out relating all the clinical outcome variables to the duration of seizures, and the Youden index was used to define the optimal cut-off of duration to determine the self-limiting of the studied seizures. Table 1 shows the cut-off values for each clinical outcome with the respective value of the area under the ROC curve (AUC) and confidence interval at 95%. The AUC is greater in the analysis involving as clinical outcomes EEG alterations at 6 months (0,89; CI 95%: 0,84-0,94) and EEG alterations at 12 months (0,82; CI 95%: 0,68-0,95). In these cases, the cut-off value corresponds to 6 min, so this value was used for subsequent analysis (Table 1). In Table 2 is summarized normal EEG and normal clinical at 6 and 12 months follow-up and normal follow-up at 12 months (no seizures recurrence) for both group with p-value (Table 2) (Fig. 1a-c), Median, mean, minimum, and maximum duration of seizures in both groups are in Table 3 (Table 3). In total, patients with seizure < 6 min was 178 (83%), with seizure > 6 min was 37 patients (17%). 22 patient on 37 (59,4%) have rescue medication, first therapy (time in min) in ≥ 6 min group has been administrated with a median (IQ range) of 4, 5 (3-5), a mean (SD) of 4, 14 (1,08) and a min-max of 2-6. The drugs used to stop seizure were: endorectal diazepam in in the group < 6 min and 11 (29, 7%) in the group ≥ 6 min. During 12-month follow-up, the number of FSs was 5 in the group < 6 min (2, 8%) and 8 (21, 6%) in the group ≥ 6 min. Table 2 shows differences between the two groups obtaining a p-value through the exact Fisher test. Patients with FSs greater than 6 min have a higher probability of presenting EEG alterations at 6 and 12 months and of seizures occurrence the following year ( Table 2). The neurological objectivity is abnormal in when the duration of seizure is greater than 6 min.

Discussion
Currently, the diagnostic criteria for SFSs state a cut-off of 15 min to define the benignity, furthermore, the seizure should be self-limiting, tonic-clonic generalized, with normal EEG, neuroimaging and lab results [13]. SFSs occurs due to prolonged inflammatory status altering neuronal susceptibility and hyperactivity [14][15][16][17]. A cut-off of 15 min of seizure may result in a delay in the administration of the life-saving anticonvulsant drug with potential complications. Studies on animal models showed that prolonged seizures lead to brain injury in paralyzed and mechanically ventilated baboons [15]. The last guidelines of management of the convulsive crisis indicate to intervene within the first 2-3 min of the beginning of the crisis [13]. It has been seen that in the major majority of cases SFSs halt spontaneously and, those that often require the administration of the drug, have a greater risk of relapses with a worse prognosis [13]. Therefore, to date, the duration of the seizure is one of the biggest unknowns for physician regarding simple seizures. In our study, we found that most FSs ends within minutes, typically less than 5 min. Therefore, we decided to study those seizures lasting more than 5 min, evaluating what happens from the 6th minutes of seizure, by the study of EEG alterations at 6 and 12 months, and the neurological clinical signs during follow-up. We decided to identify the optimal cut-off to determine the benignity of simple febrile seizures, and based on our statistical analysis (Youden index) we found this cut-off at 6 min. After 6 min of duration, we found a higher probability of presenting EEG alterations  and abnormal neurological objectivity, such as psychomotor delay. After 6 min, the risk of recurrence of FSs during the following year is higher. Moreover, literature data showed that the occurrence of multiple FSs is also associated with a slight but statistically significant increase in the risk of subsequent epilepsy [14][15][16][17][18][19] In our study, only 2% of the studied children developed epilepsy, in particular, children who had a seizure with a duration greater than 6 min. These were the only children who developed epilepsy.
According to our study, the treatment of febrile seizures must take place within 6 min of the start of the critical episode. Within this cut-off, it appears prudent to intervene pharmacologically, not being able to predict its duration and knowing that, beyond this limit, physicians could be faced with a potential complex crisis or epilepticus status with a greater risk of recurrence and a negative neurodevelopmental prognosis.

Limits of the study
One of the limitations of our study is that it is a retrospective study and, therefore, burdened with potential bias. The duration of the seizures can be underestimated by inaccurate emergency room records even in very good institutions. The EEG study was performed outside the two hospital centers. It was not prescribed by our neuropediatrics, but it was requested by the General Practitioner at the specific request of the parents. In fact, most of the parents who come to our observation with their children for FS have faced this disease as a dramatic event, feeling the need to further diagnose any neurological problems by studying the EEG.

Conclusions
In conclusion, we suggest to assign a new cut-off of 6 min to determine the benignity of a FS, reducing the standardized cut-off of 15 min. This cut-off is important because when a child arrives at the emergency room with a FS of fewer minutes (less than 6 min), physicians can reassure parents about the benign prognosis. A FS lasting more than 6 min should be evaluated for neurological follow-up, making the parents aware that the seizure could recur and that their child is likely to undergo an antiepileptic therapy.
Everything defined as SFS can take place according to our study in less than 6 min. Therefore, we propose introducing a new entity: typical FS, characterized by self-limiting, generalized tonic-clonic seizure, lasting < 6 min, with normal EEG and imaging, occasional recurrence (once/twice in 1 year, once in a lifetime, does not recur within the same illness) in a child of typical age, with no neurological disorder. Other studies are mandatory to thoroughly test this proposal.