document3

Febrile seizuresSearch date August 2007Leena D Mewasingh

ABSTRACTINTRODUCTION: Simple febrile seizures are generalised in onset, last less than 15 minutes, and do not occur more than once in 24 hours.Complex seizures are longer lasting, have focal symptoms, and can recur within 24 hours.This review only deals with simple febrile seizures.About 25% of children in the USA and Western Europe, and 69% of infants and children in Japan, will have experienced at least onefebrile seizure by the age of 5 years. Simple febrile seizures may slightly increase the risk of developing epilepsy, but have no known adverseeffects on behaviour, scholastic performance, or neurocognition. METHODS AND OUTCOMES: We conducted a systematic review andaimed to answer the following clinical questions: What are the effects of treatments given during episodes of fever in children with one ormore previous simple febrile seizures? What are the effects of long-term (daily, for more than 1 month) anticonvulsant treatment in childrenwith a history of simple febrile seizures? What are the effects of treatments on reducing the risk of subsequent epilepsy in children with ahistory of simple febrile seizures? We searched: Medline, Embase, The Cochrane Library and other important databases up to August 2007(BMJ Clinical Evidence reviews are updated periodically, please check our website for the most up-to-date version of this review).We includedharms alerts from relevant organisations such as the US Food and Drug Administration (FDA) and the UK Medicines and Healthcare productsRegulatory Agency (MHRA). RESULTS: We found 19 systematic reviews, RCTs, or observational studies that met our inclusion criteria.We performed a GRADE evaluation of the quality of evidence for interventions. CONCLUSIONS: In this systematic review we present infor-mation relating to the effectiveness and safety of the following interventions: anticonvulsants (intermittent or continuous), and antipyretictreatments (physical antipyretic measures, paracetamol, ibuprofen).

QUESTIONSWhat are the effects of treatments given during episodes of fever in children with one or more previous simplefebrile seizures?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3What are the effects of long-term (daily, longer than 1 month) anticonvulsant treatment in children with a historyof simple febrile seizures?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5What are the effects of treatments on reducing the risk of subsequent epilepsy in children with a history of simplefebrile seizures?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

INTERVENTIONSTREATING EPISODES OF FEVER

Unknown effectivenessAntipyretic treatments (physical antipyretic measures,paracetamol, ibuprofen) . . . . . . . . . . . . . . . . . . . . . . 3

Likely to be ineffective or harmfulAnticonvulsants (intermittent) . . . . . . . . . . . . . . . . . . 4

LONG-TERM CONTINUOUS ANTICONVULSANTPROPHYLAXIS

Trade-off between benefits and harmsAnticonvulsants (continuous) . . . . . . . . . . . . . . . . . . 5

PREVENTING EPILEPSY

Unlikely to be beneficialAnticonvulsants (intermittent and continuous) . . . . . 7

Key Points

Febrile seizures are defined as events in infancy or childhood usually occurring between 3 months and 5 years ofage associated with a fever, but without evidence of intracranial infection, or defined cause for the seizure.

Simple febrile seizures are generalised in onset, last less than 15 minutes, and do not occur more than once in24 hours. Complex seizures are longer lasting, have focal symptoms, and can recur within 24 hours. This reviewonly deals with simple febrile seizures.About 25% of children in the USA and Western Europe, and 69% of infants and children in Japan will haveexperienced at least one febrile seizure by age 5 years.Simple febrile seizures may slightly increase the risk of developing epilepsy, but have no known adverse effectson behaviour, scholastic performance, or neurocognition.

We do not know whether antipyretics are useful in treating episodes of fever to prevent seizure recurrence inchildren with one or more previous simple febrile seizures.

Intermittent anticonvulsants used in treating episodes of fever to prevent seizure recurrence in children are as-sociated with adverse effects, including hyperactivity, irritability, and difficulties with speech, activity level, orsleep.

Continuous anticonvulsant treatment may be effective for reducing recurrence in children with a history of simplefebrile seizures, but is associated with adverse effects: for example, phenobarbital is associated with cognitiveimpairments and behavioural adverse effects, including hyperactivity, irritability, and aggressiveness.

Child health

BMJ Publishing Group Ltd 2008. All rights reserved. . . . . . . . . . . . . . . . . . . . . . 1 . . . . . . . . . . . . . . . . . . . . . Clinical Evidence 2008;05:324

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Anticonvulsants do not appear to reduce the risk of epilepsy up to 12 years later in children with a history ofsimple febrile seizures.

DEFINITION Febrile seizures are divided into three types: simple febrile seizures, complex febrile seizures, andfebrile status epilepticus. This review focuses on children with simple febrile seizures. TheNational Institutes of Health (NIH) definition of a febrile seizure is "an event in infancy or childhoodusually occurring between 3 months and 5 years of age associated with a fever, but without evidenceof intracranial infection or defined cause for their seizure", [1] after having excluded children withprevious afebrile seizures. Another definition from the International League Against Epilepsy (ILAE)is that of "a seizure occurring in childhood after 1 month of age associated with a febrile illness notcaused by an infection of the central nervous system (CNS), without previous neonatal seizuresor a previous unprovoked seizure, and not meeting the criteria for other acute symptomatic seizures".[2]

In working practice, the lower age limit for febrile seizures is generally taken to be 6 months,given concerns regarding the possibility of an underlying serious but treatable infection in youngerinfants masquerading as a febrile seizure (e.g. meningitis). A simple febrile seizure is a generalisedseizure, often tonic-clonic, lasting less than 15 minutes in duration, which does not occur morethan once in 24 hours, and is followed by full recovery within 1 hour.Treatment for the actual seizureis generally not indicated, given the short duration. In over 80% of children the duration of the febrileseizure is less than 10 minutes, and in only about 9% of children do they last longer than 15 minutes.[3]

Often, by the time the child presents to hospital, the seizure has already stopped. A febrileseizure may also be the presenting sign of a fever episode. This review does not include childrenexperiencing complex febrile seizures, which are characterised by any of the following features:greater than 15 minutes in duration, focal symptoms, recurrence within 24 hours, and not followedby full consciousness within 1 hour. Investigations including neuroimaging and lumbar punctureare often warranted. Also excluded from this review are children experiencing febrile statusepilepticus, which lasts longer than 30 minutes and requires treatment. Addressing parental anx-iety forms a key part of the management of simple febrile seizures, as parents' (unspoken) worrywith a first seizure is that their child might have died. However, there is little in the medical literatureabout this aspect of education and reassurance in management of simple febrile seizures.

INCIDENCE/PREVALENCE

About 25% of children in the USA and Western Europe, and 69% of infants and children in Japan,will have experienced at least one febrile seizure, simple or complex, by the age of 5 years. Else-where the incidence varies, being 510% in India, and as high as 14% in Guam. [2] There are nospecific data available for simple febrile seizures.

AETIOLOGY/RISK FACTORS

While the exact cause of simple febrile seizures is unknown, it is thought to be multifactorial, withboth genetic and environmental factors having been shown to contribute to its pathogenesis. [4][5] Increasingly, a genetic predisposition is recognised, with febrile seizures occurring in families.However, the exact mode of inheritance is not known, and seems to vary between families. Whilepolygenic inheritance is likely, there is a small number of families identified with an autosomaldominant pattern of inheritance of febrile seizures, leading to the description of a "febrile seizuresusceptibility trait" with an autosomal dominant pattern of inheritance with reduced penetrance. [6][4]

In addition, mutations in several genes have been found that account for enhanced susceptibil-ity to febrile seizures. [5] [7] [8] [9] [10] [11] A familial epilepsy syndrome exists (GeneralisedEpilepsy with Febrile Seizures Plus [GEFS+]), in which patients can have classical febrile seizures,febrile seizures that persist beyond 5 years (hence FS+), and/or epilepsy. Similar genetic factorshave been identified that are involved in both febrile seizures and GEFS+. [5] [12] Although theexact molecular mechanisms of febrile seizures are yet to be understood, underlying mutationshave been found in genes encoding the sodium channel and the gamma amino-butyric acid A re-ceptor. [12] [13] Both of these channels are also associated with another early epilepsy syndrome,Severe Myoclonic Epilepsy of Infancy (SMEI) which often begins with prolonged febrile seizure(either complex febrile seizure or febrile status) with subsequent seizures precipitated by fever. [12]With regards risk factors, febrile seizures are more frequent in children attending day-care centres,and in those with a first- or second-degree relative with a history of febrile seizures. [14] The riskof another child having febrile seizures is one in five if one sibling is affected, and one in three ifboth parents and a previous child have had febrile seizures. [15] Other risk factors associated withan increased rate of febrile seizure recurrence include young age at onset (less than 12 months),history of simple or complex febrile seizures, and body temperature at onset of less than 40 C.[14] [16]

Among these, age at onset seems the most constant predictive factor, with 50% of childrenaged less than 12 months, and 30% of children aged more than 12 months, presenting with a re-current febrile seizure. Positive family history of epilepsy is not consistently associated with increasedsimple febrile seizure recurrence. [16]

PROGNOSIS Simple febrile seizures may slightly increase the risk of developing epilepsy, [17] but have no adverseeffects on behaviour, scholastic performance, or neurocognition. The risk of developing epilepsy

BMJ Publishing Group Ltd 2008. All rights reserved. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Febrile seizuresChild health

is increased further in children with a history of complex febrile seizures. [18] [19] [20] [21] A strongassociation exists between febrile status epilepticus or febrile seizures characterised by focalsymptoms and later development of temporal lobe epilepsy. [17] [22]

AIMS OFINTERVENTION

To reduce febrile seizures and prevent recurrence, and to prevent the development of epilepsy,with minimal adverse effects.

OUTCOMES Febrile seizures; development of epilepsy; adverse effects of treatment.

METHODS BMJ Clinical Evidence search and appraisal August 2007. The following databases were used toidentify studies for this review: Medline 1966 to August 2007, Embase 1980 to August 2007, andThe Cochrane Database of Systematic Reviews and Cochrane Central Register of ControlledClinical Trials 2007, Issue 3. Additional searches were carried out using these websites: NHSCentre for Reviews and Dissemination (CRD) for Database of Abstracts of Reviews of Effects(DARE) and Health Technology Assessment (HTA), Turning Research into Practice (TRIP), andNICE. Abstracts of the studies retrieved from the initial search were assessed by an informationspecialist. Selected studies were then sent to the author for additional assessment, using pre-de-termined criteria to identify relevant studies. Study design criteria for evaluation in this chapterwere: published systematic reviews and RCTs in any language, and containing more than 20 indi-viduals of whom more than 80% were followed up. There was no minimum length of follow-up re-quired to include studies. We included all studies described as blinded, open, open label, ornot blinded. We also searched for cohort studies on specific harms of named interventions. In ad-dition, we use a regular surveillance protocol to capture harms alerts from organisations such asthe FDA and the UK Medicines and Healthcare products Regulatory Agency (MHRA), which areadded to the review as required. Some systematic reviews meta-analysed RCTs of children withboth simple febrile seizures and complex febrile seizures, and in some RCTs the type of febrileseizure is unspecified; we have reported this throughout the text. We have performed a GRADEevaluation of the quality of evidence for interventions included in this review (see table, p 10 ).

QUESTION What are the effects of treatments given during episodes of fever in children with one ormore previous simple febrile seizures?

OPTION ANTIPYRETIC TREATMENTS (PHYSICAL ANTIPYRETIC MEASURES, PARACETAMOL,IBUPROFEN). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Recurrence of febrile seizuresAntipyretic drugs compared with placebo Ibuprofen or paracetamol may be no more effective at 12 years at reducingthe proportion of children (who have previously had 1 simple febrile seizure) with recurrence of febrile seizures ( verylow-quality evidence ).NoteWe found no direct information about physical methods of temperature reduction in the treatment of children withsimple febrile seizures.

For GRADE evaluation of interventions for febrile seizures, see table, p 10 .

Benefits: Physical methods of temperature reduction:We found no systematic review or RCTs in children with febrile seizures (see comment below).Antipyretic drugs versus placebo:We found one systematic review (search date 2003), which identified two RCTs. [23] The first RCT(230 children with 1 previous simple febrile seizure) found no significant difference in the proportionof children who had recurrence of febrile seizures at 1 year between ibuprofen given during episodesof fever and placebo (31/111 [28%] with ibuprofen v 36/119 [30%] with placebo; reported as notsignificant). The second RCT (180 children with 1 previous simple febrile seizure) compared fourinterventions: paracetamol plus placebo, paracetamol plus diazepam, diazepam plus placebo, andplacebo plus placebo. It found no significant difference in the number of recurrent febrile seizuresat 2 years between paracetamol given during episodes of fever and placebo (5.2% with paracetamolplus placebo v 8.2% with placebo plus placebo; reported as not significant). [23] The review hadweak methods; it used inadequate search methods that are difficult to replicate, and had no specificinclusion/exclusion criteria. [23]

Harms: Physical methods of temperature reduction:We found no RCTs in children with previous simple febrile seizures (see comment below).

BMJ Publishing Group Ltd 2008. All rights reserved. ............................................................ 3


Antipyretic drugs versus placebo:The systematic review gave no information on adverse effects. [23] We found one systematic review(search date 2004, 3 RCTs, 254 people) comparing paracetamol versus placebo in children withfever. [24] It found no significant difference in the proportion of children who had adverse effectsbetween paracetamol and placebo (9/130 [7%] with paracetamol v 4/124 [3%] with placebo; RR1.84, 95% CI 0.65 to 5.18).

Comment: Physical methods of temperature reduction:We found one systematic review (search date 2004), which identified two RCTs (120 children)comparing tepid sponging versus paracetamol in children with an axillary temperature of38.540.0 C with UTI, malaria, or both. One of the RCTs (80 children) assessed febrile seizures.[24]

It found no significant difference in febrile seizures at 2 hours between sponging and paracetamol(1/40 [3%] with tepid sponging v 0/40 [0%] with paracetamol; RR 0.33, 95% CI 0.01 to 7.95). Thereview found no significant difference in adverse effects, including shivering, goose pimples, anddiscomfort, between sponging and paracetamol (6/55 [11%] with tepid sponging v 2/65 [3%] withparacetamol; RR 0.26, 95% CI 0.07 to1.01). As this review excluded children with previous simplefebrile seizures, we have not reported it in the previous section. [24]

OPTION ANTICONVULSANTS (INTERMITTENT). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Recurrence of febrile seizuresCompared with placebo We don't know whether intermittent diazepam given to children during a febrile episode ismore effective at reducing the risk of febrile seizure recurrence in children with a history of simple or complex febrileseizures ( very low-quality evidence ). Clobazam given during a febrile episode may be more effective at reducingthe risk of recurrence of febrile seizures (very low-quality evidence).NoteDiazepam has been associated with increased hyperactivity, lethargy, irritability, and with difficulties with speech,activity level, or sleep. We found no clinically important results about intermittent compared with continuous anticon-vulsants for treating children with febrile seizures.


Benefits: We found two systematic reviews [25] [26] and two subsequent RCTs [27] [28] assessing intermittentanticonvulsants (diazepam and clobazam) given during a febrile episode to reduce recurrence inchildren with a history of simple or complex febrile seizures.

Intermittent anticonvulsants versus placebo or no treatment:The first review (search date not reported, 4 RCTs, 791 children) compared intermittent diazepamversus placebo. [25] It found that diazepam given during a febrile episode significantly reduced theproportion of children with febrile seizure recurrence compared with placebo (44/393 [11%] withintermittent diazepam v 68/398 [17%] with placebo; OR for recurrence 0.60, 95% CI 0.40 to 0.90).The second review (search date not reported) identified three of the same RCTs as the first sys-tematic review, and one additional RCT comparing intermittent diazepam versus no treatment(1060 children). [26] It found no significant difference in febrile seizure recurrence at up to 24 monthsbetween diazepam given during a febrile episode and placebo or no treatment (92/537 [17%] withintermittent diazepam v 134/522 [26%] with placebo or no treatment; P = 0.20; see comment below).[26] The first subsequent RCT (40 children with 1 or more episodes of febrile seizure) comparedclobazam given during episodes of fever versus placebo. [27] The children had 108 episodes offever over a mean of 9.9 months; 60 episodes were treated with clobazam and 48 with placebo.The RCT found that clobazam given during a febrile episode significantly reduced the rate of seizurerecurrence compared with placebo (6/48 [12%] episodes with placebo v 1/60 [2%] episodes withclobazam; P = 0.01).The RCT randomised children and analysed episodes of fever; it is unclearwhether adjustments were made to allow for this, but it is likely that results would remain significanteven with adjustment. The second subsequent RCT (60 children who completed the study, aged6 months to 5 years, presenting with 1 or more episodes of febrile seizure) also compared clobazamgiven during febrile episodes versus placebo. [28] The children had 312 episodes of fever over aperiod of 6 months; 151 episodes were treated with clobazam and 161 with placebo. In both groups,parents were advised to treat temperatures above 38 C with antipyretic (paracetamol) and coldsponging. The RCT found that clobazam given during a febrile episode significantly reduced therecurrence of seizures compared with placebo (recurrence of seizures with febrile episodes: 9/30[30%] people with clobazam v 25/30 [83%] people with placebo; P less than 0.001). In this RCT,each arm was also treated with antipyretic measures as described above, so that the significantresults in the treatment group would appear to be attributable to clobazam only. However, the RCTdid not report the method of randomisation used, and stated that the analysis was not by intentionto treat; in addition, the number of children lost to follow-up after randomisation into the RCT is notclear. [28]



Intermittent anticonvulsants versus continuous anticonvulsants:We found no systematic review or RCTs comparing intermittent anticonvulsants (diazepam) versuscontinuous anticonvulsants (phenobarbital or sodium valproate).

Harms: Intermittent anticonvulsants versus placebo or no treatment:The reviews gave no information on adverse effects. [25] [26] However, when we assessed individ-ual papers, two RCTs included in the reviews reported information on adverse events. [29] [30] Thefirst RCT (185 children with simple or complex febrile seizures) found that diazepam significantlyincreased the number of days that children were hyperactive (defined as agitation and inability tokeep still) compared with placebo (138 days with diazepam v 34 days with placebo; P less than0.0003). [29] The second RCT found that 59/153 (39%) children taking intermittent diazepam hadadverse effects, including: ataxia; lethargy; irritability; or difficulties with speech, activity level, orsleep. One child taking placebo had a rash. [30] In the first subsequent RCT, clobazam significantlyincreased ataxia compared with placebo (5 [8.3%] with clobazam v 0 [0%] with placebo; P = 0.04),while the second subsequent RCT reported that "no significant difference in adverse drug-reactionprofile was observed except irritability, which was occasionally more in clobazam group" (irritability:4/30 [13%] with clobazam v 1/30 [3%] with placebo; further details and statistical analysis betweengroups not reported). [27] [28] We found one further prospective controlled study (139 children whoentered the study after their first febrile seizure) which compared intermittent rectal diazepam givenduring the first 2 days of a febrile illness (68 children) with a group that received no prophylaxisduring fever (71 children). [31] Treatments were allocated by odd and even dates (further detailsnot reported), and follow-up was 3 years. Children with complex febrile seizures (approximately19%) and febrile status epilepticus (approximately 3%) were also included.The study reported thatadverse effects with diazepam were mild and transient, and no long-term side effects wererecorded during the 3-year follow-up (no further numerical data or statistical analysis of adverseeffects between groups reported). [31]

Intermittent anticonvulsants versus continuous anticonvulsants:We found no RCTs.

Comment: Intermittent anticonvulsants versus placebo or no treatment:The two reviews identified three of the same RCTs. [25] [26] However, one of these RCTs [30] hasbeen reported differently in each review. In the first systematic review [25] the recurrence rates forthis RCT are reported as 7/202 (3.5%) in children taking diazepam and 29/204 (14.2%) in childrentaking placebo; [25] in the second systematic review, they are reported as 37/202 (18.3%) in childrentaking diazepam compared with 53/204 (30%) in children taking placebo. [26] This may explainhow reviews with predominantly the same included RCTs came to different conclusions.The mode,dose, and frequency of administration of diazepam varied in each RCT. Most of the RCTs identifiedby the reviews had weak methods. The first RCT was small. In the second RCT, 50 children (25%)taking diazepam and 55 children (27%) taking placebo were lost to follow-up.The third RCT reportedpoor compliance in children taking diazepam, which was significantly different from those takingplacebo. [25] [26]

QUESTION What are the effects of long-term (daily, longer than 1 month) anticonvulsant treatment inchildren with a history of simple febrile seizures?

OPTION ANTICONVULSANTS (CONTINUOUS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Recurrence of febrile seizuresContinuous phenobarbital compared with placebo/no treatment Phenobarbital may be more effective at reducingfebrile seizure recurrence in children with a history of simple or complex febrile seizures ( low-quality evidence ).Continuous sodium valproate compared with placebo/no treatment Continuous sodium valproate may be no moreeffective at reducing febrile seizure recurrence in children with a history of simple or complex febrile seizures (low-quality evidence).Continuous anticonvulsants compared with each other We dont know whether phenobarbital is more effective thansodium valproate at reducing the proportion of children with febrile seizure recurrence (low-quality evidence).NotePhenobarbital may be associated with cognitive impairment, and with behavioural problems including hyperactivity,irritability, and aggression. Serious adverse events which may be associated with sodium valproate include hepato-toxicity and haematological toxicity, both of which may occasionally be fatal.




Benefits: We found two systematic reviews, which between them identified eight RCTs assessing continuousanticonvulsants given during a febrile episode in children with a history of simple or complex febrileseizures. [25] [26]

Continuous phenobarbital versus placebo or no treatment:The first review (search date not reported) identified six RCTs (598 children) comparing continuousphenobarbital versus placebo. [25] It found that continuous phenobarbital significantly reduced theproportion of children with febrile seizure recurrence compared with placebo (71/290 [24%] withphenobarbital v 114/308 [37%] with placebo; OR 0.54, 95% CI 0.38 to 0.76; NNT 17, 95% CI 10to 85). However, there was significant statistical heterogeneity among the trials (P less than 0.01).[27] The second review (search date not reported, 8 RCTs, 6 included in the previous review, 975children) also found that phenobarbital significantly reduced the proportion of children with febrileseizure recurrence compared with placebo or no treatment (90/483 [19%] with phenobarbital v184/492 [37%] with placebo or no treatment; RR 0.51, 95% CI 0.32 to 0.82; P less than 0.01). [26]The second review also found heterogeneity among the trials (figures not reported; see commentbelow).Continuous sodium valproate versus placebo or no treatment:We found one systematic review (search date not reported), which identified three RCTs (278children) comparing sodium valproate versus placebo or no treatment. [26] It found no significantdifference between groups in the proportion of children with febrile seizure recurrence (29/102[28%] with sodium valproate v 34/114 [30%] with placebo or no treatment; RR 0.74, 95% CI 0.24to 2.23, P = 0.59; calculated by random-effects model). [26] The authors of the review suggest that,if only the small (48 children), placebo-controlled RCT is considered, there is a significant decreasein recurrent febrile seizures with sodium valproate compared with placebo (1/22 [4%] with sodiumvalproate v 9/26 [35%] with placebo; RR 0.13, 95% CI 0.02 to 0.96, P = 0.01).Continuous anticonvulsants versus each other:We found one systematic review (search date not reported), [25] which identified one RCT (69children with 1 previous febrile seizure, type not reported) comparing three interventions: dailysodium valproate, daily phenobarbital, and placebo. [32] It found no significant difference in theproportion of children with febrile seizure recurrence between phenobarbital and sodium valproate(4/21 [19%] with phenobarbital v 1/22 [4%] with sodium valproate; reported as not significant, Pvalue not reported).

Harms: Continuous phenobarbital versus placebo or no treatment:The reviews gave no information on adverse effects. [25] [26] However, when we assessed individ-ual papers, five RCTs identified by the review included information on adverse effects. [33] [34] [35][36] [37] We also found one additional RCT assessing cognitive and behavioural adverse effects.[38]

All of the RCTs found that phenobarbital increased cognitive and behavioural adverse effectscompared with placebo ( see table 1, p 9 ).Continuous sodium valproate versus placebo or no treatment:The systematic review gave no information on adverse effects. [26] Known rare, serious adverseeffects of sodium valproate include hepatotoxicity and haematological toxicity. [39] [40] [41] Althoughvalproate hepatotoxicity may be dose dependent, it can, more rarely, be an idiosyncratic phe-nomenon which means that it is often irreversible and difficult to predict on the basis of labora-tory monitoring. [39] Blood disturbances can also be dose dependent, with direct bone marrowsuppression leading to aplastic anaemia or peripheral cytopenia affecting one or more cell lines,or even fatal bone marrow failure. [41]

Continuous anticonvulsants versus each other:The review [25] and RCT [32] gave no information on adverse effects.

Comment: Searches in the first review included studies in Spanish, and English language RCTs; [25] searchesin the second review were restricted to English language. [26] Inclusion criteria were stated in bothreviews. Methods for data extraction were also stated in both reviews, [25] [26] albeit more brieflyin one [25] than the other. Doses of phenobarbital or sodium valproate differed across RCTs. MostRCTs did not assess compliance and, although the adverse effects associated with anticonvulsantsare well known, potential adverse events associated with their use were not routinely monitoredor investigated. Despite this lack of formal monitoring of adverse effects, the authors of the individ-ual RCTs and the reviews often concluded that the benefit : harm ratios for phenobarbital andsodium valproate were unfavourable.

Heterogeneity among RCTs identified:Tests for statistical heterogeneity were carried out in both reviews; both found significant hetero-geneity among the phenobarbital and sodium valproate RCTs. [25] [26] However, the authors of



the first review concluded that the effectiveness of phenobarbital could not be shown. [25] This isnot quite in keeping with the findings of a significantly lower recurrence of febrile seizures in childrentaking phenobarbital. [25] The second review found that, even taking the heterogeneity of the trialsof phenobarbital into account, the overall effect remained positive, with inconsistencies amongRCTs being ascribed rather to the degree of seizure preventing effect. [26]

QUESTION What are the effects of treatments on reducing the risk of subsequent epilepsy in childrenwith a history of simple febrile seizures?

OPTION ANTICONVULSANTS (INTERMITTENT AND CONTINUOUS). . . . . . . . . . . . . . . . . . . . . . . . . . .

Incidence of epilepsyIntermittent diazepam compared with no prophylaxis Intermittent diazepam given during a febrile episode may beno more effective at reducing the incidence of epilepsy at 12 years in children with one previous simple or complexfebrile seizure ( low-quality evidence ).Continuous Phenobarbital (daily/intermittent) compared with no treatment Phenobarbital given daily or intermittentlymay be no more effective at reducing the proportion of children who develop epilepsy at 6.3 years (low-quality evi-dence).For GRADE evaluation of interventions for febrile seizures, see table, p 10 .

Benefits: We found no systematic review of anticonvulsants for preventing the development of epilepsy inchildren with simple febrile seizures.

Intermittent diazepam:We found one quasi-randomised RCT (289 children with one previous simple or complex febrileseizure), comparing intermittent diazepam (given during a febrile episode) versus no prophylaxis(diazepam only in case of febrile seizure) that followed up the children at 12 years after the initialstudy had finished. [42] It found no significant difference between groups in the incidence ofepilepsy at 12 years (0.8% with intermittent diazepam v 0.7 % with diazepam during seizure; re-ported as not significant, P value not reported). The RCT found that the risk of having epilepsy atthe mean age of 14 years was 1/250 (0.4%) after simple febrile seizures and 1/40 (2.5%) aftercomplex febrile seizures. [42] There was no significant difference in the incidence of epilepsy withdifferent types of seizure (reported as not significant, P value not reported).The quasi-randomisedRCT assigned groups alternatively, depending on whether the child was admitted on an odd oreven date.

Continuous phenobarbital:We found one RCT (355 children with 1 previous simple or complex febrile seizure) comparingthree interventions: daily phenobarbital, intermittent phenobarbital (given at the onset of fever),and no treatment. [43] It found that, although continuous phenobarbital significantly reduced recurrentfebrile seizures compared with no treatment, there was no significant difference in the proportionof children who developed epilepsy over a mean 6.3 years (7/116 [6%] with daily phenobarbital v6/158 [4%] with intermittent phenobarbital v 1/126 [1%] with no treatment; reported as not significant,P value not reported).Continuous sodium valproate:We found no systematic review or RCTs about the effects of sodium valproate on preventingepilepsy in children with simple febrile seizures.

Harms: Intermittent diazepam:The quasi-randomised RCT found no significant difference in full scale, verbal, or performance in-telligence quotients (IQ; measured by the Wechsler Intelligence Scale for Children [WISC] generalintelligence test), memory, reading tests, and overall scholastic performance at 12 years betweenintermittent diazepam and diazepam during seizures (absolute results tabulated; P value not sig-nificant for all outcomes). [42]

Continuous phenobarbital:The RCT gave no information on adverse effects. [43]

Continuous sodium valproate:We found no RCTs.

Comment: The RCT assessing children who had taken phenobarbital for febrile seizures suggested that childrenwho subsequently developed epilepsy differed from those who did not in the frequency of neonatalabnormality, family history of mental retardation, focal initial febrile seizures, and delay in psychomo-



tor milestones before the initial febrile seizure. It also suggested that half of the children who devel-oped mental retardation had histories of delay in psychomotor milestones before the initial febrileseizure. [43]

GLOSSARYLow-quality evidence Further research is very likely to have an important impact on our confidence in the estimateof effect and is likely to change the estimate.Very low-quality evidence Any estimate of effect is very uncertain.

SUBSTANTIVE CHANGESAnticonvulsants (intermittent) One small RCT (60 children) comparing clobazam versus placebo added to thetwo systematic reviews already reported, and benefits and harms data enhanced. [28] One prospective study (139children) comparing diazepam versus placebo added to the harms section, and harms data enhanced. [31] Categori-sation of 'Anticonvulsants (intermittent)' unchanged (Likely to be ineffective or harmful).

REFERENCES1. Anonymous. Consensus statement. Febrile seizures: long-term management of

children with fever-associated seizures. Pediatrics 1980;66:10091012.2. Commission on Epidemiology and Prognosis. International League Against

Epilepsy. Guidelines for epidemiologic studies on epilepsy. Epilepsia1993;34:592595. [PubMed]

3. Berg AT, Shinnar S, Berg AT, et al. Unprovoked seizures in children with febrileseizures: short-term outcome. Neurology 1996;47:562568.

4. Offringa M, Bossuyt PM, Lubsen J, et al. Risk factors for seizure recurrence inchildren with febrile seizures: a pooled analysis of individual patient data fromfive studies. J Pediatr 1994;124:574584.

5. Audenaert D, Van Broeckhoven C, De Jonghe P, et al. Genes and loci involvedin febrile seizures and related epilepsy syndromes. [Review] [69 refs]. Hum Mut2006;27:391401.

6. Iwasaki N, Nakayama J, Hamano K, et al. Molecular genetics of febrile seizures.Epilepsia 2002;43:3235.

7. Johnson EW, Dubovsky J, Rich SS, et al. Evidence for a novel gene for familialfebrile convulsions, FEB2, linked to chromosome 19p in an extended family fromthe Midwest. Hum Mol Genet 1998;7:6367. [PubMed]

8. Gerard F, Pereira S, Robaglia-Schlupp A, et al. Clinical and genetic analysis ofa new multigenerational pedigree with GEFS+ (Generalized Epilepsy with FebrileSeizures Plus). Epilepsia 2002;43:581586. [PubMed]

9. Nabbout R, Prud'homme JF, Herman A, et al. A locus for simple pure febrileseizures maps to chromosome 6q22-q24. Brain 2002;125:26682680. [PubMed]

10. Kananura C, Haug K, Sander T, et al. A splice-site mutation in GABRG2 associ-ated with childhood absence epilepsy and febrile convulsions. Arch Neurol2002;59:11371141. [PubMed]

11. Baulac S, Gourfinkel-An I, Nabbout R, et al. Fever, genes, and epilepsy. LancetNeurol 2004;3:421430. [PubMed]

12. Hirose S, Mohney RP, Okada M, et al.The genetics of febrile seizures and relatedepilepsy syndromes.Brain Dev 2003;25:304312.

13. Nakayama J, Arinami T, Nakayama Junko, et al. Molecular genetics of febrileseizures.Epilepsy Research 2006;70 Suppl 1:S190S198.

14. Knudsen FU. Febrile seizures treatment and outcome. Brain Dev1996;18:438449. [PubMed]

15. Doose H, Maurer A. Seizure risk in offspring of individuals with a history of febrileconvulsions. Eur J Pediatr 1997;156:476481. [PubMed]

16. Berg AT, Shinnar S, Hauser WA, et al. Predictors of recurrent febrile seizures:a metaanalytic review. J Pediatr 1990;116:329337. [PubMed]

17. Verity CM, Golding J. Risk of epilepsy after febrile convulsions: a national cohortstudy. BMJ 1991;303:13731376. [Erratum in: BMJ 1992;304:147]

18. Berg AT. Febrile seizures and epilepsy: the contributions of epidemiology. Pae-diatr Perinat Epidemiol 1992;6:145152. [PubMed]

19. Verity CM, Butler NR, Golding J. Febrile convulsions in a national cohort followedup from birth. I Prevalence and recurrence in the first five years of life. BMJ1985;290:13071310. [PubMed]

20. Verity CM, Butler NR, Golding J. Febrile convulsions in a national cohort followedup from birth. II Medical history and intellectual ability at 5 years of age. BMJ1985;290: 13111315. [PubMed]

21. Verity CM, Greenwood R, Golding J. Long-term intellectual and behavioral out-comes of children with febrile convulsions. N Engl J Med 1998;338:17231738.[PubMed]

22. Tarkka R, Paakko E, Pyhitinen J, et al. Febrile seizures and mesial temporalsclerosis: no association in a long-term follow-up study. Neurology2003;60:215218. [PubMed]

23. El-Radhi AS, Barry W. Do antipyretics prevent febrile convulsions? Arch Dis Child2003;88:641642. Search date 2003; primary sources Pubmed, The CochraneLibrary, and the Database of Abstracts of Reviews of Effects (Dare). [PubMed]

24. Meremikwu M, Oyo-Ita A. Paracetamol for treating fever in children. In: TheCochrane Library: Issue 3, 2007. Chichester, UK: John Wiley & Sons, Ltd. Search

date 2004; primary sources Cochrane Infectious Diseases Group SpecializedRegister, Central, Medline, Embase, Lilacs, Science Citation Index, referencelists of articles, and contact with researchers in the field.

25. Masuko AH, Castro AA, Santos GR, et al. Intermittent diazepam and continuousphenobarbital to treat recurrence of febrile seizures: a systematic review withmeta-analysis. Arq Neuropsiquiatr 2003;61:897901. Search date not reported;primary sources Cochrane Center of Brazil, Medline, Lilacs, Embase, handsearches of references of articles retrieved, thesis indexed at the Biblioteca Re-gional Medicina/Panamerican Health Organization of the World Health Organiza-tion, abstracts sent to medical meetings, and letters to experts. [PubMed]

26. Temkin NR. Antiepileptogenesis and seizure prevention trials with antiepilepticdrugs: meta-analysis of controlled trials. Epilepsia 2001;42:515524. Searchdate not reported, primary sources Medline, Embase, Cochrane Controlled TrialsRegister, hand searches of references of articles retrieved, attendance at meet-ings, and other information communication. [PubMed]

27. Rose W, Kirubakaran C, Scott JX. Intermittent clobazam therapy in febrileseizures. Indian J Pediatr 2005;72:3133. [PubMed]

28. Bajaj AS, Bajaj BK, Puri V, et al. Intermittent clobazam in febrile seizures: AnIndian experience. J Pediatr Neurol 2005;3:1923.

29. Autret E, Billard C, Bertrand P, et al. Double-blind, randomized trial of diazepamversus placebo for prevention of recurrence of febrile seizures. J Pediatr1990;117:490494. [PubMed]

30. Rosman NP, Colton T, Labazzo J, et al. A controlled trial of diazepam adminis-tered during febrile illnesses to prevent recurrence of febrile seizures. N Engl JMed 1993;329:7984. [PubMed]

31. Pavlidou E, Tzitiridou M, Panteliadis C. Effectiveness of intermittent diazepamprophylaxis in febrile seizures: long-term prospective controlled study. J ChildNeurol 2006;21:10361040.

32. Mamelle N, Mamelle JC, Plasse JC, et al. Prevention of recurrent febrile convul-sions a randomized therapeutic assay: sodium valproate, phenobarbitone andplacebo. Neuropediatrics 1984;15:3742.

33. Farwell JR, Lee YJ, Hirtz DG, et al. Phenobarbital for febrile seizures effectson intelligence and on seizure recurrence. N Engl J Med 1990;322:364369.[Erratum in: N Engl J Med 1992;326:144]

34. Thilothammal N, Kannan, Krishnamurthy PV, et al. Role of phenobarbitone inpreventing recurrence of febrile convulsions. Indian Pediatr 1993;30:637642.[PubMed]

35. Bacon CJ, Hierons AM, Mucklow JC, et al. Placebo-controlled study of phenobar-bitone and phenytoin in the prophylaxis of febrile convulsions. Lancet1981;2:600604. [PubMed]

36. Camfield PR, Camfield CS, Shapiro SHL, et al.The first febrile seizure antipyret-ic instruction plus either phenobarbital or placebo to prevent recurrence. J Pediatr1980;97:1621. [PubMed]

37. Wolf SM, Forsythe A. Behavior disturbance, phenobarbital, and febrile seizures.Pediatrics 1978;61:728731. [PubMed]

38. Camfield CS, Chaplin S, Doyle AB, et al. Side effects of phenobarbital in toddlers:behavioral and cognitive aspects. J Pediatr 1979;95:361365. [PubMed]

39. Fenichel GM, Greene HL. Valproate hepatotoxicity: two new cases, a summaryof others, and recommendations. Pediatr Neurol 1985;1:109113. [PubMed]

40. Konig SA, Elger CE, Vassella F, et al. Recommendations for blood studies andclinical monitoring in early detection of valproate-associated liver failure. Resultsof a consensus conference 9 May 11 May 1997 in Berlin. Nervenarzt1998;69:835840. [In German] [PubMed]

41. Acharya S, Bussel JB. Hematologic toxicity of sodium valproate. J PediatrHematol Oncol 2000;22:6265. [PubMed]

42. Knudsen FU, Paerregaard A, Andersen R, et al. Long term outcome of prophy-laxis for febrile convulsions. Arch Dis Child 1996;74:1318. [PubMed]

43. Wolf SM, Forsythe A. Epilepsy and mental retardation following febrile seizuresin childhood. Acta Paediatr Scand 1989;78:291295. [PubMed]

Leena D MewasinghConsultant Paediatric Neurologist

Imperial College Healthcare NHS TrustLondon

UK



Competing interests: LDM has been reimbursed by the pharmaceutical companies JanssenCilag and UCB Pharma for attending epilepsy congresses and epilepsy meetings.

TABLE 1 Adverse effects reported in RCTs comparing phenobarbital (PHB) versus placebo or notreatment in children with at least one previous simple febrile seizure.

Adverse eventsPopulationRefNegative effect on cognition: 2-year follow-up mean IQ, PHB v placebo: 8.4points, 95% CI 13.3 points to 3.5 points; P = 0.0057). 6 months after weaningand discontinuation of PHB, mean IQ PHB v placebo: 5.2 points, 95% CI 10.5points to 0.04 points; P = 0.052)

217 children with at least 1 previoussimple febrile seizure

[33]

Adverse effects necessitating withdrawal: 3/60 (5%) PHB-treated children hadintolerable adverse effects (defined as effects persistent for longer than 1 month),including hyperkinetic behaviour, extreme irritability, fussiness, aggressiveness, allof whom withdrew from the study owing to the adverse effects. 1/30 (3.3%) of chil-dren taking placebo withdrew for unknown reasons

90 children with 2 or more previoussimple febrile seizures (60 takingPHB v 30 taking placebo)

[34]

Negative effects on behaviour: Many parents of children taking PHB reporteddeterioration in behaviour, as did many parents of children taking placebo (absolutenumbers and P value not reported); 20% reported slight improvement in some as-pects of behaviour when PHB was withdrawn

138 children with 1 previous simplefebrile seizure

[35]

Adverse effects necessitating withdrawal: 4/39 (10%) in both groups withdrewbecause of intolerable adverse effects


[36]

Negative effects on behaviour: 46/109 (42%) children taking continuous PHBdeveloped a behaviour disorder, usually hyperactivity, compared with 22/120 (18%)having no treatment. Hyperactivity spontaneously disappeared in 52% of children.Continuous PHB was prematurely discontinued in 25/46 (54%) of the children withbehaviour abnormality (20% of those treated)

371 children with 1 previous simplefebrile seizure (109 taking continuousPHB v 142 intermittent PHB v 120 notreatment)

[37]

Negative effects on behaviour (increased fussiness and sleep disturbanceclassed as transient, dose related, or unacceptable): transient: 8/35 (23%) withPHB v 7/30 (23%) with placebo; dose related: 4/35 (11%) with PHB v 0/30 (0%)with placebo; unacceptable: 3/35 (9%) with PHB v 1/30 (3%) with placebo.Decreased comprehension: Children taking PHB had lower scores on memoryconcentration items on the StanfordBinet Intelligence scale at 8- to 12-month follow-up compared with children taking placebo, although the difference between groupswas not significant (absolute numbers not reported; P = 0.07).


[38]

IQ, intelligence quotients; PHB, phenobarbital; Ref, reference.



TABLE GRADE evaluation of interventions for febrile seizures

Febrile seizures, development of epilepsy, adverse effectsImportantoutcomes

CommentGRADEEf-fectsize

Direct-ness

Con-sisten-cy

Quali-ty

Typeof evi-dence

ComparisonOutcomeNumber ofstudies(partici-pants)What are the effects of treatments given during episodes of fever in children with one or more previous simple febrile seizures?

Quality points deducted forincomplete reporting of re-sults and methodologicalweaknesses

Verylow

00034Antipyretics vplacebo

Recurrenceof febrileseizures

2 (410) [23]

Quality points deducted forweak methodologies. Consis-tency point deducted forconflicting results. Direct-ness points deducted for dif-ferences in doses, methodsof administration, and compli-ance

Verylow

02114Intermittent anti-convulsants (di-azepam) v place-bo/no treatment


5 (1850)[25] [26]

Quality points deducted forsparse data, and method-ological flaws (no intention-to-treat- analysis, not report-ing method of randomisa-tion, and uncertainty aboutfollow-up)

Verylow

00034Intermittent anti-convulsants(clobazam) vplacebo/no treat-ment


2 (100)[27] [28]

What are the effects of long-term (daily, longer than 1 month) anticonvulsant treatment in children with a history of simple febrile seizures?Consistency point deductedfor heterogeneity amongstudies. Directness pointdeducted for differences indoses used

Low01104Continuous phe-nobarbital vplacebo/no treat-ment


8(1573) [25][26]

Consistency point deductedfor heterogeneity amongstudies. Directness pointdeducted for differences indoses used

Low01104Continuous sodi-um valporate vplacebo/no treat-ment


3 (216)[25]

Quality points deducted forsparse data and incompletereporting of results

Low00024Continuous anti-convulsants veach other


1 (43) [32]

What are the effects of treatments on reducing the risk of subsequent epilepsy in children with a history of simple febrile seizures?Quality points deducted forquasi-randomisation and in-complete reporting of results

Low00024Intermittent di-azepam v no pro-phylaxis

Incidence ofepilepsy

1 (290)[42]

Quality point deducted forincomplete reporting of re-sults. Directness point de-ducted for underlying differ-ences in risk of developingepilepsy

Low01014Phenobarbital(daily or intermit-tent) v no treat-ment

Incidence ofepilepsy

1(400) [43]

Type of evidence: 4 = RCT; 2 = Observational; 1 = Non-analytical/expert opinion. Consistency: similarity of results across studiesDirectness: generalisability of population or outcomesEffect size: based on relative risk or odds ratio

Disclaimer

The information contained in this publication is intended for medical professionals. Categories presented in Clinical Evidence indicate ajudgement about the strength of the evidence available to our contributors prior to publication and the relevant importance of benefit andharms. We rely on our contributors to confirm the accuracy of the information presented and to adhere to describe accepted practices.Readers should be aware that professionals in the field may have different opinions. Because of this and regular advances in medical researchwe strongly recommend that readers' independently verify specified treatments and drugs including manufacturers' guidance. Also, thecategories do not indicate whether a particular treatment is generally appropriate or whether it is suitable for a particular individual. Ultimatelyit is the readers' responsibility to make their own professional judgements, so to appropriately advise and treat their patients.To the fullest extent permitted by law, BMJ Publishing Group Limited and its editors are not responsible for any losses, injury or damagecaused to any person or property (including under contract, by negligence, products liability or otherwise) whether they be direct or indirect,special, incidental or consequential, resulting from the application of the information in this publication.

BMJ Publishing Group Ltd 2008. All rights reserved. ........................................................... 10


document3

Documents