June 1999, VOLUME103 /ISSUE 6

Technical Report: Treatment of the Child With Simple Febrile Seizures

  1. Robert J. Baumann, MD


Overview. Simple febrile seizures that occur in children ages 6 months to 5 years are common events with few adverse outcomes. Those who advocate therapy for this disorder have been concerned that such seizures lead to additional febrile seizures, to epilepsy, and perhaps even to brain injury. Moreover, they note the potential for such seizures to cause parental anxiety. We examined the literature to determine whether there was demonstrable benefit to the treatment of simple febrile seizures and whether such benefits exceeded the potential side effects and risks of therapy. The therapeutic approaches considered included continuous anticonvulsant therapies, intermittent therapy, or no anticonvulsant therapy.

Methods. This analysis focused on the neurologically healthy child between 6 months and 5 years of age whose seizure is brief (<15 minutes), generalized, and occurs only once during a 24-hour period during a fever. Children whose seizures are attributable to a central nervous system infection and those who have had a previous afebrile seizure or central nervous system abnormality were excluded. A review of the current literature was conducted using articles obtained through searches in MEDLINE and additional databases. Articles were obtained following defined criteria and data abstracted using a standardized literature review form. Abstracted data were summarized into evidence tables (Tables 1 through 7).

Results. Epidemiologic studies demonstrate a high risk of recurrent febrile seizures but a low, though increased, risk of epilepsy. Other adverse outcomes either don't occur or occur so infrequently that their presence is not convincingly demonstrated by the available studies. Although daily anticonvulsant therapy with phenobarbital or valproic acid is effective in decreasing recurrent febrile seizures, the risks and potential side effects of these medications outweigh this benefit. No medication has been shown to prevent the future onset of recurrent afebrile seizures (epilepsy). The use of intermittent diazepam with fever after an initial febrile seizure is likely to decrease the risk of another febrile seizure, but the rate of side effects is high although most families find the perceived benefits to be low. Although antipyretic therapy has other benefits, it does not prevent additional simple febrile seizures.

Conclusions. The Febrile Seizures Subcommittee of the American Academy of Pediatrics' Committee on Quality Improvement used the results of this analysis to derive evidence-based recommendations for the treatment of simple febrile seizures. The outcomes anticipated as a result of the analysis and development of the practice guideline include: 1) to optimize practitioner understanding of the scientific basis for using or avoiding various proposed treatments for children with simple febrile seizures; 2) to improve the health of children with simple febrile seizures by avoiding therapies with high potential for side effects and no demonstrated ability to improve children's eventual outcomes; 3) to reduce costs by avoiding therapies that will not demonstrably improve children's long-term outcomes; and 4) to help the practitioner educate caregivers about the low risks associated with simple febrile seizures.

  • febrile seizures
  • epilepsy
  • valproic acid
  • carbamazepine
  • phenytoin
  • diazepam
  • phenobarbital
  • sodium valproate
  • pyridoxine
  • Abbreviation:
    NCPP =
    National Collaborative Perinatal Project
  • The debate over whether children with recurrent febrile seizures benefit from anticonvulsant therapy began early in this century.1 An important advance was the identification of the subgroup of children with simple febrile seizures; a subgroup that is large, remarkably homogeneous, and healthy at 7- and 10-year follow-ups.2 ,3 Furthermore, the recognition of such favorable outcomes has accentuated the need to balance the risk of any treatment with an expected benefit. Epidemiologic studies helped to identify this subgroup, demonstrated their predominantly favorable outcomes, and confirmed what has long been known: febrile seizures are common events. Of youngsters in a British birth cohort, 2.7% had febrile seizures, 88% of whom had simple febrile seizures.3 ,4


    This parameter is limited to children with simple febrile seizures defined as neurologically healthy infants and children between 6 months and 5 years of age whose seizure is brief (<15 minutes), generalized, and occurs only once during a 24-hour period in a febrile child. This definition is easily applied in the usual clinical circumstances and has the additional advantages of encompassing most children with febrile seizures and defining a relatively homogeneous group of patients.4 This practice parameter excludes children whose seizures are attributable to a central nervous system infection (symptomatic febrile seizures) and those who have had a previous afebrile seizure or central nervous system abnormality (secondary febrile seizures).


    Proponents of therapy for simple febrile seizures have worried that repeated simple febrile seizures will lead to more febrile seizures and possibly to afebrile seizures (epilepsy). They also have been apprehensive that these seizures will cause brain injury and thus diminish intelligence or impair motor coordination.5

    A child who has experienced a single simple febrile seizure is likely to experience another. As epidemiologic data indicate, this recurrence rate is strongly age-related. The younger the child at the time of the first event, the more likely there will be subsequent events. In the National Collaborative Perinatal Project (NCPP), half the subjects with onset of febrile seizures during the first year versus approximately 30% with onset after the first year had one or more additional febrile seizures.2 This project included 1706 prospectively studied children with febrile seizures from approximately 54 000 pregnancies.

    The risk of experiencing a single afebrile seizure or two or more afebrile seizures (defined as epilepsy) is elevated when comparing children with simple febrile seizures with the general age-matched population.3 ,6 ,7 In the NCPP, the risk factors for epilepsy after a febrile seizure were a positive family history of afebrile seizures, preexisting neurologic abnormality, and a complicated initial febrile seizure. Interestingly, the age at first febrile seizure and the number of febrile seizures did not alter this risk.6 At age 7 years, only 1.9% of children with simple febrile seizures and negative family histories of epilepsy had experienced a single afebrile seizure, and epilepsy had developed in 0.9%. The comparable figures for study children who never experienced a febrile seizure were 0.9% for a single afebrile seizure and 0.5% for epilepsy.2 ,6 Similar rates were seen in the large British cohort study that included all surviving neonates born in the United Kingdom during 1 week in April 1970. These children were followed until age 10 years, and 305 had an initial simple febrile seizure. Of the 305 children, 8 (2.6%) subsequently had an afebrile seizure and epilepsy eventually developed in 5 (1.6%). The comparable number with epilepsy among the 14 278 children who never had febrile seizures was 53 (0.4%).3

    Although the risk of epilepsy among children with simple febrile seizures is elevated, the rate is still low, and the number of children in any given study is small. These numbers provide some understanding of the difficulty of designing a population-based study to determine if any treatment for the prevention of simple febrile seizures would subsequently prevent the development of epilepsy.

    Investigators have attempted to look at this issue. At the Kaiser Foundation Hospitals in Southern California, Wolf and colleagues8 ,9 studied 400 children who had febrile seizures (identified from lumbar puncture reports). They divided the children into three study groups: those who received phenobarbital daily, those who received phenobarbital only with fever, and those who received no therapy. Follow-up lasted a mean of 6.3 years. No difference was found in the rate of afebrile seizures. This study included children with complex as well as simple febrile seizures, many children did not receive the prescribed medication (approximately one third), and the study was small and could have missed a statistically valid effect.

    In another study, 289 children with febrile seizures were randomized to rectal diazepam prophylaxis at the onset of any fever or rectal diazepam therapy during any febrile seizure. At age 14 years, there was no difference in intelligence, coordination, or occurrence of epilepsy between the two groups. The number of study patients was small, there are questions regarding compliance, and patients with complex and simple febrile seizures were included.10

    Evidence for adverse outcomes other than epilepsy has been sought. The NCPP had the benefit of longitudinal examination of a predefined group of children. No evidence of death in relation to asymptomatic febrile seizures was found, and examination of the children revealed no evidence for the development of motor deficits.2 There also has been concern about cognitive deficits in relation to febrile seizures. The NCPP found no effect on intelligence among 431 children with febrile seizures who were compared with their siblings.2 Comparisons of children with simple febrile seizures with the general population also have found no adverse effect.11 ,12 Smith and Wallace13 believed that they found an adverse effect of repeated febrile seizures on intelligence as measured by the Griffith Mental Development Scale. Because they studied children with simple and complex febrile seizures, it is possible that underlying neurologic disease predisposed to further seizures and to lower scores on retesting.


    Pertinent articles previously obtained by a MEDLINE search and a search of the Epilepsy Foundation of America database4were reviewed and supplemented by references suggested by members of the Committee and the Committee's consultants. More than 300 articles were reviewed.

    The goal of the review was to identify articles that met the following criteria:

    • The study children had simple febrile seizures that were convincingly differentiated from afebrile seizures and other types of febrile seizures.

    • The subjects with simple febrile seizures were reasonably representative of children with simple febrile seizures.4 ,5

    • A suitable control group was included in the study. Preference was given to blinded protocols.



    There are several studies in which phenobarbital administered daily successfully prevented recurrent febrile seizures. Camfield and associates14 randomized 79 children who had had a first simple febrile seizure to receive phenobarbital at 5 mg/kg per day in a single dose or a placebo. Compliance was monitored by use of the urine fluorescence of a riboflavin additive and by measurements of serum phenobarbital levels. There was a significant difference in the incidence of recurrent febrile seizures between the phenobarbital recipients (2/39 [5%]) and the placebo group (10/40 [25%]). Neither parents nor investigators knew which subjects received the active drug.14 Investigators found no significant difference in IQ (using Stanford-Binet or Bayley Scales) between the placebo and phenobarbital groups after 8 to 12 months of therapy. Nevertheless, phenobarbital was demonstrated to decrease memory and concentration in proportion to higher serum phenobarbital levels. Transient sleep disturbances and daytime fussiness were more common among phenobarbital recipients, but by 1 year, the two groups were indistinguishable. This was partially accounted for by 4 children receiving phenobarbital whose side effects resolved after the dosage was reduced.

    In a controlled trial comparing phenobarbital (5 mg/kg per day) with phenytoin (8 mg/kg per day) and placebo, Bacon and associates15 also found phenobarbital to be effective. In younger children, the febrile seizure recurrence rate was 9% (2/22) for phenobarbital recipients versus 44% (12/27) placebo recipients. This trial included subjects with complicated febrile seizures who were stratified proportionally into the three groups. The study had major problems with compliance. All phenobarbital-treated children with a recurrence for whom drug levels were obtained at the time of recurrence had a plasma level <15 mg/L. Interestingly the reported behavioral changes were similar in the subjects treated with phenobarbital and a placebo.

    Mamelle et al16 compared phenobarbital (3 to 4 mg/kg per day), valproate (30 to 40 mg/kg per day in 2 doses), and placebo in a randomized single-blind study of infants with a first simple febrile seizure. They found significantly fewer recurrences in the valproate (1/22 [4.5%]) and phenobarbital (4/21 [19%]) groups compared with the placebo group (9/26 [35%]). Compliance was measured by serum drug levels. Only 5 subjects were removed from therapy because of side effects; all were described as having agitation and all were receiving phenobarbital. Other studies, some with designs that were less rigorous, also found phenobarbital to be effective,17–20including the previously mentioned Kaiser Foundation study.8

    Not all studies have found phenobarbital to be effective. Heckmatt et al21 found recurrent febrile seizures in 14 (19%) of 73 control subjects, 10 (11%) of 88 children for whom phenobarbital was prescribed, and 4 (8%) of 49 who actually took the prescribed phenobarbital (4 to 5 mg/kg per day in divided doses). These last 4 subjects had plasma phenobarbital levels >16 mg/L at the time of recurrence. Although the differences between the treatment groups are not statistically important, they seem to suggest that an effect favoring phenobarbital might have been evident had the numbers been larger or the duration of the study longer.

    Children who had complicated febrile seizures analyzed by intention to treat experienced no difference in recurrence rate between phenobarbital-treated subjects and controls.22 The study described a poor rate of compliance and seemed to show that a medication is not effective if parents are unable to administer it. Early in the study when compliance was high, 56% of phenobarbital recipients (4 to 5 mg/kg once per day) and 35% of placebo recipients were reported to have side effects. The study found that the mean IQ was 8.4 points lower in the phenobarbital group than in the placebo group (95% CI: −13.3–−3.5, P = .0057) at the end of the 2-year study, with an IQ differential that persisted 6 months after the taper of medication had begun. The analysis of these data was complicated by the low compliance rates, the fact that 24 (26%) of 94 placebo recipients and 53 (64%) of 83 phenobarbital recipients were prescribed phenobarbital after the study ended, and the inclusion of subjects with complicated febrile seizures.22

    Phenobarbital is associated with impairment of short-term memory and concentration and worsening of behavior.23–25 Most data on the effects of phenobarbital have been obtained from adults or from children with epilepsy.23 ,24 The drug's effect seems most prominent at the onset of therapy.23 The reported effect in children with simple febrile seizures varies among studies. In the study by Camfield et al,14 parents were only aware of side effects early in the study, but higher serum levels were associated with decreased memory concentration. Smith and Wallace13found no effect of therapy but believed that repeated seizures in children with complicated febrile seizures were associated with lower mental development scores. Wolf and Forsythe26 reported hyperactivity in 46 (42%) of 109 children treated with phenobarbital (initial dose, 3 to 4 mg/kg per day, adjusted to give a serum level of 10 to 15 μg/mL) for febrile seizures compared with 21 (17.5%) of 120 not receiving phenobarbital. As in other studies, a substantial rate of improvement was noted in both groups with time. When 25 children from each group were extensively tested, no cognitive differences could be detected.27

    Valproic Acid

    A number of studies have demonstrated the effectiveness of this agent in preventing recurrent febrile seizures.18 ,19 The study by Mamelle et al16 typifies the studies that found valproic acid to be more effective than phenobarbital.28Although no severe adverse effects are described among the children participating in the febrile seizure trials, the numbers in these trials are small. Valproic acid therapy is associated with fatal hepatotoxicity,29 ,30 pancreatitis,31 renal toxicity,32 hematopoietic disturbances,33 and other problems.


    Carbamazepine was not effective for febrile seizures in preliminary trials and, thus, has not been studied widely.34 ,35 In a double-blind trial of carbamazepine (20 mg/kg per day in twice daily doses) vs phenobarbital (4 to 5 mg/kg per day) involving children with complicated febrile seizures, Antony and Hawke17 reported recurrent febrile seizures in 9 (47%) of 19 carbamazepine recipients and 2 (10%) of 21 phenobarbital recipients.


    As with carbamazepine, preliminary studies showed no evidence that phenytoin was effective for febrile seizures, so it has not been studied extensively.36 In a randomized, controlled study of children with simple and complex febrile seizures, the recurrence rate in the phenytoin group (8 mg/kg per day) of younger children was 33% (9/27) compared with 9% (2/22) for the phenobarbital group (5 mg/kg per day) and 44% (12/27) for the equivalent placebo group.15


    Antipyretic Agents

    Because simple febrile seizures occur only in conjunction with a fever, it has seemed logical to try to prevent these seizures by using aggressive antipyretic therapy. In the randomized, double-blind study by Camfield and associates,14 all subjects received detailed instruction about temperature control, including antipyretic use with any rectal temperature higher than 37.2°C (99°F). Ten (25%) of 40 subjects using only temperature control had recurrences compared with 2 (5%) of 39 receiving continuous phenobarbital. A randomized, controlled trial using a complicated study design with placebo, low-dose diazepam, and acetaminophen also found no evidence that acetaminophen prevented recurrent febrile seizures.37 In this protocol, the diazepam-treated children who had previously experienced a febrile seizure received a rectal diazepam solution (if they weighed <7 kg, they received 2.5 mg; if 7 to 15 kg, 5 mg; and if >15 kg, 10 mg) followed in 6 hours by 0.2 mg/kg three times a day whenever they were febrile. The antipyretic treatment group received 10 mg/kg of acetaminophen four times per day.

    In children hospitalized after a simple febrile seizure, Schnaiderman et al38 found that acetaminophen (15 to 20 mg/kg per dose) given every 4 hours did not prevent a second febrile seizure during that admission any better than giving acetaminophen sporadically. The two groups also had the same frequency, duration, and height of temperature elevations. There is no evidence that aggressive antipyretic therapy prevents recurrent febrile seizures.


    The use of intermittent diazepam prophylaxis for febrile seizures is well-reported in the literature.39 Autret and colleagues,40 in a randomized, controlled multicenter study, found that oral diazepam (0.5 mg/kg initially, then 0.2 mg/kg every 12 hours) was no more effective than a placebo in preventing recurrent febrile seizures. Most of the children had simple febrile seizures, and the data were analyzed by intention to treat. Recurrence was experienced by 15 (16%) of 93 children in the diazepam group compared with 18 (20%) of 92 children in the placebo group. Although parents “were instructed verbally, in writing, and by demonstration,” there were major problems with compliance. In children with recurrences, only 1 (7%) of/15 diazepam recipients and 7 (39%) of 18 placebo recipients received the medication or placebo as prescribed. The difference between these two groups is significant. The reasons that the subjects did not receive their assigned treatment included the following: 1) 7 in each group had a seizure as the first sign of illness, 2) 5 parents in the diazepam group and 4 in the placebo group did not give the medication, and 3) 2 children in the diazepam group would not take their medication. Because 14 (93%) of the 15 children for whom diazepam was prescribed who had a recurrence had not received their prescribed medication, these data demonstrate that a treatment is not effective if parents cannot or will not administer it before the febrile seizure occurs. The only noted side effect was hyperactivity, which was significantly more frequent in the diazepam group (138 vs 34 days).

    By contrast, in a similarly well-designed, randomized, double-blind, placebo-controlled trial, Rosman et al41 found oral diazepam to be significantly more effective than placebo when analyzed by intention to treat. A 44% reduction in the risk of febrile seizures per person-year occurred with diazepam. Children in the diazepam group had 675 febrile episodes and 41 febrile seizures, of which 7 occurred while receiving the study medication. Comparable figures for the placebo group were 526 febrile episodes and 72 febrile seizures, of which 38 occurred while receiving the placebo. These investigators describe febrile seizures as “highly upsetting” to the parent population, which may have influenced adherence. Not surprisingly, they found that a higher rate of side effects accompanied their subjects' better compliance. Of the diazepam recipients, 59 (39%) had at least one “moderate” side effect and a similar number had a “mild” side effect.

    NOTE FOR TABLES: The main drugs listed in the table titles also may have been administered in combination with other drugs.


    • The recommendations in this statement do not indicate an exclusive course of treatment or serve as a standard of medical care. Variations, taking into account individual circumstances, may be appropriate.




    1. BU1.
    2. BU2.
    3. BU3.
    4. BU4.
    5. BU5.
    6. BU6.
    7. BU7.
    8. BU8.
    9. BU9.
    10. BU10.
    11. BU11.
    12. BU12.
    13. BU13.

    Phenobarbital—Side Effects

    1. BU14.
    2. BU15.
    3. BU16.
    4. BU17.
    5. BU18.
    6. BU19.
    7. BU20.


    1. BU21.
    2. BU22.
    3. BU23.
    4. BU24.
    5. BU25.
    6. BU26.
    7. BU27.
    8. BU28.

    Sodium Valproate

    1. BU29.
    2. BU30.
    3. BU31.
    4. BU32.
    5. BU33.
    6. BU34.
    7. BU35.


    1. BU36.
    2. BU37.


    1. BU38.


    1. BU39.

    Antipyretic Measures

    1. BU40.
    2. BU41.
    3. BU42.
    4. BU43.
    5. BU44.
    6. BU45.