Prophylactic drug management for febrile seizures in children

Martin Offringa; Richard Newton; Sarah J Nevitt; Katerina Vraka

doi:10.1002/14651858.CD003031.pub4

. 2021 Jun 16;2021(6):CD003031. doi: 10.1002/14651858.CD003031.pub4

Prophylactic drug management for febrile seizures in children

Martin Offringa ^1,^✉, Richard Newton ², Sarah J Nevitt ³, Katerina Vraka ⁴

Editor: Cochrane Epilepsy Group

PMCID: PMC8207248 PMID: 34131913

Abstract

Background

Febrile seizures occurring in a child older than one month during an episode of fever affect 2‐4% of children in Great Britain and the United States and recur in 30%. Rapid‐acting antiepileptics and antipyretics given during subsequent fever episodes have been used to avoid the adverse effects of continuous antiepileptic drugs.

This is an updated version of a Cochrane Review previously published in 2017.

Objectives

To evaluate primarily the effectiveness and safety of antiepileptic and antipyretic drugs used prophylactically to treat children with febrile seizures; and also to evaluate any other drug intervention where there is a sound biological rationale for its use.

Search methods

For the latest update we searched the following databases on 3 February 2020: Cochrane Register of Studies (CRS Web), MEDLINE (Ovid, 1946 to 31 January 2020). CRS Web includes randomised or quasi‐randomised controlled trials from PubMed, Embase, ClinicalTrials.gov, the World Health Organization International Clinical Trials Registry Platform (ICTRP), the Cochrane Central Register of Controlled Trials (CENTRAL), and the specialised registers of Cochrane Review Groups including the Cochrane Epilepsy Group. We imposed no language restrictions and contacted researchers to identify continuing or unpublished studies.

Selection criteria

Trials using randomised or quasi‐randomised participant allocation that compared the use of antiepileptics, antipyretics or recognised Central Nervous System active agents with each other, placebo, or no treatment.

Data collection and analysis

For the original review, two review authors independently applied predefined criteria to select trials for inclusion and extracted the predefined relevant data, recording methods for randomisation, blinding, and exclusions. For the 2016 update, a third review author checked all original inclusions, data analyses, and updated the search. For the 2020 update, one review author updated the search and performed the data analysis following a peer‐review process with the original review authors. We assessed seizure recurrence at 6, 12, 18, 24, 36, 48 months, and where data were available at age 5 to 6 years along with recorded adverse effects. We evaluated the presence of publication bias using funnel plots.

Main results

We included 42 articles describing 32 randomised trials, with 4431 randomised participants used in the analysis of this review. We analysed 15 interventions of continuous or intermittent prophylaxis and their control treatments. Methodological quality was moderate to poor in most studies. We found no significant benefit for intermittent phenobarbital, phenytoin, valproate, pyridoxine, ibuprofen, or zinc sulfate versus placebo or no treatment; nor for diclofenac versus placebo followed by ibuprofen, paracetamol, or placebo; nor for continuous phenobarbital versus diazepam, intermittent rectal diazepam versus intermittent valproate, or oral diazepam versus clobazam.

There was a significant reduction of recurrent febrile seizures with intermittent diazepam versus placebo or no treatment at six months (risk ratio (RR) 0.64, 95% confidence interval (CI) 0.48 to 0.85; 6 studies, 1151 participants; moderate‐certainty evidence), 12 months (RR 0.69, 95% CI 0.56 to 0.84; 8 studies, 1416 participants; moderate‐certainty evidence), 18 months (RR 0.37, 95% CI 0.23 to 0.60; 1 study, 289 participants; low‐certainty evidence), 24 months (RR 0.73, 95% CI 0.56 to 0.95; 4 studies, 739 participants; high‐certainty evidence), 36 months (RR 0.58, 95% CI 0.40 to 0.85; 1 study, 139 participants; low‐certainty evidence), 48 months (RR 0.36, 95% CI 0.15 to 0.89; 1 study, 110 participants; moderate‐certainty evidence), with no benefit at 60 to 72 months (RR 0.08, 95% CI 0.00 to 1.31; 1 study, 60 participants; very low‐certainty evidence).

Phenobarbital versus placebo or no treatment reduced seizures at six months (RR 0.59, 95% CI 0.42 to 0.83; 6 studies, 833 participants; moderate‐certainty evidence), 12 months (RR 0.54, 95% CI 0.42 to 0.70; 7 studies, 807 participants; low‐certainty evidence), and 24 months (RR 0.69, 95% CI 0.53 to 0.89; 3 studies, 533 participants; moderate‐certainty evidence), but not at 18 months (RR 0.77, 95% CI 0.56 to 1.05; 2 studies, 264 participants) or 60 to 72 months follow‐up (RR 1.50, 95% CI 0.61 to 3.69; 1 study, 60 participants; very low‐certainty evidence).

Intermittent clobazam compared to placebo at six months resulted in a RR of 0.36 (95% CI 0.20 to 0.64; 1 study, 60 participants; low‐certainty evidence), an effect found against an extremely high (83.3%) recurrence rate in the controls, a result that needs replication.

When compared to intermittent diazepam, intermittent oral melatonin did not significantly reduce seizures at six months (RR 0.45, 95% CI 0.18 to 1.15; 1 study, 60 participants; very‐low certainty evidence).

When compared to placebo, intermittent oral levetiracetam significantly reduced recurrent seizures at 12 months (RR 0.27, 95% CI 0.15 to 0.52; 1 study, 115 participants; very low‐certainty evidence).

The recording of adverse effects was variable. Two studies reported lower comprehension scores in phenobarbital‐treated children. Adverse effects were recorded in up to 30% of children in the phenobarbital‐treated groups and 36% in benzodiazepine‐treated groups. We found evidence of publication bias in the meta‐analyses of comparisons for phenobarbital versus placebo (seven studies) at 12 months but not at six months (six studies); and valproate versus placebo (four studies) at 12 months. There were too few studies to identify publication bias for the other comparisons.

The methodological quality of most of the included studies was low or very low. Methods of randomisation and allocation concealment often did not meet current standards, and 'treatment versus no treatment' was more commonly seen than 'treatment versus placebo', leading to obvious risks of bias.

Authors' conclusions

We found reduced recurrence rates for intermittent diazepam and continuous phenobarbital, with adverse effects in up to 30% of children. The apparent benefit for clobazam treatment in one trial needs to be replicated. Levetiracetam also shows benefit with a good safety profile; however, further study is required. Given the benign nature of recurrent febrile seizures, and the high prevalence of adverse effects of these drugs, parents and families should be supported with adequate contact details of medical services and information on recurrence, first aid management, and, most importantly, the benign nature of the phenomenon.

Plain language summary

Prophylactic drug management for febrile seizures in children

Background

Seizures occurring with a fever (febrile seizures) in children are common and affect about one in 30 children under the age of six years. On average, one out of three children who have had a febrile seizure will have at least one more. We reviewed the evidence about the effect of drugs to prevent seizures (antiepileptics), drugs to lower temperature (antipyretics), and zinc in children with febrile seizures.

Objective

We wanted to know in how many children these drugs would prevent a recurrence of febrile seizures or cause unwanted effects.

Methods

We included 32 studies with a total of 4431 children in the review. Children who had had at least one febrile seizure were assigned to one of two or more treatment groups. The studies recorded any further seizures at various time intervals between six months and up to six years of age in each group. Unwanted medication effects were also noted.

Results

The study design and evidence quality in the studies of antiepileptic drugs was often low or very low. Poor methods known to lead to obvious risks of bias were used. One issue was with the methods used to assign children to study groups and how random this allocation was. Other issues included whether the parents or doctors, or both, knew which group each child was in or if a treatment was compared to no treatment with no placebo (dummy pill) used. The quality of the trials of antipyretics or zinc was better, with the evidence assessed as moderate to high.

Zinc therapy was found to provide no benefit. We also found no benefit in treating children just at the time of the fever with either antipyretic drugs or most antiepileptic drugs.

A significant result was noted in some instances. For example, at times between 6 and 48 months follow‐up, intermittent diazepam (an antiepileptic drug) led to a reduction in the number of recurrent seizures by about a third. Continuous phenobarbital resulted in significantly fewer recurrences at 6, 12, and 24 months, but not at 18 and 60 to 72 months. One study showed that intermittent oral levetiracetam compared to placebo significantly reduced recurrent seizures at 12 months. When compared to intermittent diazepam, intermittent oral melatonin did not significantly reduce seizures at six months.

However, as recurrent seizures are only seen in about a third of children, this means that up to 16 children would have to be treated over a year or two to save just one child a further seizure. As febrile seizures are not harmful, we viewed these significant findings to be unimportant, in particular because adverse effects of the medications were common. Lower comprehension scores in phenobarbital‐treated children were found in two studies. In general, adverse effects were recorded in up to about a third of children in both the phenobarbital‐ and benzodiazepine‐treated groups. The benefit found for treatment with clobazam in one study published in 2011 needs to be repeated to test its reliability. Levetiracetam may be useful in treating children where family anxiety over possible seizure recurrence is high, but further study is required.

Authors' conclusions

There is currently insufficient evidence to support the use of continuous or intermittent treatment with zinc, antiepileptic or antipyretic drugs for children with febrile seizures. Febrile seizures can be frightening to witness. Parents and families should be supported with adequate contact details of medical services and information on recurrence, first aid management, and, most importantly, the benign nature of the phenomenon.

The evidence is current to February 2020.

Summary of findings

Background

This is an updated version of a Cochrane Review previously published in 2017 (Offringa 2017).

Description of the condition

The International League Against Epilepsy (ILAE) defines a febrile seizure as “a seizure occurring in childhood after one month of age associated with a febrile illness not caused by an infection of the central nervous system, without previous neonatal seizures or a previous unprovoked seizure, and not meeting the criteria for other acute symptomatic seizures” (ILEA 1993). The cumulative incidence of febrile seizures is estimated between 2% and 5% in the USA and Western Europe (Shinnar 2003; Verity 1991); between 6% to 9% in Japan; and 14% in India and Guam (ILEA 1993). Febrile seizures have a peak incidence at 18 months and are most common between the ages of six months and six years (Berg 1996; Hauser 1994; Offringa 1991).

In 2010, the ILAE proposed that febrile seizures could be organised by typical age at onset (i.e. infancy and childhood). Conventionally, febrile seizures have been classified as simple or complex based on duration, recurrence during the same illness episode, and the presence of focal features. Most febrile seizures are generalised tonic‐clonic seizures, whilst about 30% to 35% of febrile seizures have one or more complex features (focal onset, duration > 10 minutes, or multiple seizures during the illness episode) (Berg 1996). Febrile status epilepticus, a subgroup of complex febrile seizures with seizures lasting more than 30 minutes, occur in about 5% of cases (Berg 1996).

Causation is thought to be multifactorial, with environmental factors and increasing evidence for genetic factors contributing to pathogenesis (Audenaert 2006; Offringa 1994). No single susceptibility gene for febrile seizures is known. In contrast, gene identification has been successful in families with genetic epilepsies with febrile seizures plus (GEFS+) where kindreds may well include children with Dravet syndrome (Berg 2010; Kasperaviciute 2013; Tang 2013). In these conditions, febrile seizures persist beyond the age of six years; mutations have been found in sodium voltage‐gated channel alpha subunit 1 (SCN1A) and sodium voltage‐gated channel beta subunit 1 (SCN1B) (both sodium channel genes important for neurotransmission) and gamma‐aminobutyric acid type A receptor subunit gamma2 (GABRG2) (related to γ‐aminobutyric acid, an important inhibitory neurotransmitter) (Audenaert 2006; Baulac 2004; Gérard 2002; Hirose 2003; Johnson 1998; Kananura 2002; Nabbout 2002; Nakayama 2006).

Description of the intervention

Despite the frequent nature of these seizures, debate regarding the optimal management arose at an early stage (Baumann 1999), and continues. After resolution of the acute episode, the possibility of recurrent seizures during subsequent febrile illnesses must be addressed. This risk of recurrent seizures in previously healthy, untreated children was estimated in a collaborative study that used the individual data from five follow‐up studies with similar definitions of febrile seizures and risk factors (Offringa 1994). Of 2496 children with 1410 episodes of recurrent seizures in this study, 32% had at least one, 15% had at least two, and 7% had three or more recurrent seizures after a first febrile seizure. The hazard of recurrent seizures was highest between the ages of 12 and 24 months. A history of febrile or unprovoked seizures in a first‐degree family member, a relatively low temperature at the first seizure, young age at onset (< 12 months), a family history of unprovoked seizures, and a partial initial febrile seizure were all associated with an increased risk of subsequent seizures.

If a child is considered to be at increased risk of frequent or complicated seizures (Berg 1990), prophylactic medication might be considered. However, such treatment may have adverse effects on the child's behaviour and cognitive development. The decision to treat thus requires assessment of the potential risks and benefits to the child. Since 1990, at least 300 articles have been published on the drug management of seizures associated with fever (Gram 1984). This has long been a controversial area, with a persistent variety of opinions on management. Part of this controversy reflects the fact that it is uncertain whether prophylactic medication with antiepileptics and antipyretics is effective and has no important adverse effects. Yet, phenobarbital has adverse effects such as irritability, hyperactivity, and somnolence, and may even lower the cognitive development of toddlers (Farwell 1990; Herranz 1988). To avoid the side effects of continuous antiepileptic drugs (AEDs), rapid‐acting antiepileptics given only during fever periods have been used in an attempt to reduce the risk of recurrent febrile seizures. Phenobarbital at times of fever has been proven to be ineffective, probably because of the delay in achieving appropriate serum and tissue levels. Thus far, only prophylactic diazepam, given orally or rectally, has been studied in placebo‐controlled trials. The efficacy of intermittent antipyretic treatment during febrile episodes in the prevention of seizure recurrence has recently been studied.

Newton 1988 assessed the efficacy of phenobarbital and valproate for the prophylactic treatment of febrile seizures by summarising the results of all eight British placebo‐controlled clinical trials done before 1988. Data were pooled and analysed on an intention‐to‐treat basis. The overall odds ratio of recurrent febrile seizures was 0.8 for phenobarbital and 1.42 for valproate; neither result was statistically significant. The author therefore concluded that neither treatment should be recommended. A second meta‐analysis summarised four published non‐British randomised, placebo‐controlled trials conducted up to 1996 using phenobarbital as a preventive treatment of febrile seizures (Rantala 1997). The risk of recurrences was lower in children receiving continuous phenobarbital therapy than in the placebo group (odds ratio 0.54, 95% confidence interval (CI) 0.33 to 0.90). On average, eight children would have to be treated with phenobarbital for two years continuously to prevent one febrile seizure (number needed to treat for an additional beneficial outcome (NNTB) 8, 95% CI 5 to 27) (Rantala 1997).

How the intervention might work

The rationale for using prophylactic antiepileptic drugs in children with febrile seizures is to raise seizure threshold in the face of a potentially triggering fever. Antipyretics are used to attenuate the effect of fever as a triggering factor. Previous studies have shown blood and cerebrospinal fluid zinc levels to be significantly lower in children with a febrile seizure tendency than in children with afebrile seizures. Zinc level is known to stimulate the excitatory neurotransmitter glutamate and to increase the inhibitory neurotransmitter gamma‐amino‐butyric acid.

Why it is important to do this review

We undertook this review to answer the question of whether prophylactic treatment with an antiepileptic or antipyretic drug can, when compared to no therapy, decrease the likelihood of future febrile seizures in children with febrile seizures.

Objectives

Methods

Criteria for considering studies for this review

Types of studies

We included all trials using randomised or quasi‐randomised participant allocation that compared the use of antiepileptic, antipyretic agents or recognised Central Nervous System active agents with each other, or with placebo or no treatment.

Types of participants

Children aged between six months and seven years with a history of febrile seizures and who received treatment with an antiepileptic or antipyretic drug in an attempt to prevent recurrent seizures. We also planned subgroup analyses of neurologically healthy children, of children with previous recurrent seizures, and of studies limited to children at a perceived relatively high risk of recurrence.

Types of interventions

We included trials that compared one treatment with another or with placebo (or no treatment) in children with febrile seizures. Specific drugs included the benzodiazepines (diazepam, lorazepam, clobazam, and midazolam), phenytoin, phenobarbital, valproate, diclofenac, paracetamol, and ibuprofen. We planned a subgroup analysis of intermittent AED therapies versus continuous AED therapies, and of antipyretics during episodes of fever versus AED therapy during fever. A six‐month course of zinc (shown previously to have been significantly lower in children with febrile seizures) was evaluated in one study.

Types of outcome measures

Primary outcomes

Efficacy: the proportion of children with recurrence of febrile or non‐febrile seizures at certain time points after treatment onset (6 months, 12 months, 24 months, 36 months, and where data were available at age 5 to 6 years).

Secondary outcomes

Treatment adherence (as measured in the studies).
Safety: the incidence of specific adverse unwanted effects, including irritability, hyperactivity, somnolence, impaired cognitive development for phenobarbital and intermittent diazepam, gastro‐enterologic unwanted effects for valproate and antipyretics, of any administered antiepileptic or antipyretic.
As it is of clinical interest, we analysed pooled data at the chosen study time points to estimate the recurrent febrile seizure risk in the placebo and no‐treatment groups. This analysis could provide useful insight into the natural history of the disorder.

Search methods for identification of studies

Electronic searches

We ran searches for the original review in July 2008 and subsequent searches in November 2010, May 2011, November 2013, March 2015, July 2016, and October 2018. For the latest update, we searched the following databases on 3 February 2020. We imposed no language restrictions.

Cochrane Register of Studies (CRS Web), using the search strategy outlined in Appendix 1. CRS Web includes randomised or quasi‐randomised controlled trials from PubMed, Embase, the US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov, the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP), the Cochrane Central Register of Controlled Trials (CENTRAL), and the specialised registers of Cochrane Review Groups including the Cochrane Epilepsy Group.
MEDLINE (Ovid) 1946 to 31 January 2020, using the search strategy outlined in Appendix 2.

For some previous updates, Embase, Scopus, and Database of Abstracts of Reviews of Effects (DARE) were searched; however, this was no longer necessary, because relevant studies in Embase and Scopus are included in CRS Web, and DARE is a closed archive.

Searching other resources

We checked the reference lists of articles identified by the electronic searches for additional studies. We also contacted researchers in the field to identify any ongoing or unpublished studies.

Data collection and analysis

Selection of studies

For this review update, three review authors (RN, KV, and MO) independently assessed trials for inclusion, resolving any disagreements by discussion.

Data extraction and management

For this review update, three review authors (RN, KV, and MO) extracted the outcome data specified above as well as the following data, resolving any disagreements by discussion. For the previous update of this review another review author (Martinus Cozijnsen) checked all the extracted data.

Methodological and trial design

Method of randomisation.
Method of double‐blinding.
Whether any participants had been excluded from the reported analyses.

Where data were missing, we attempted to contact trial authors for this information.

Participant and demographic information

Total number of participants allocated to each treatment group or audited in any protocol.
The proportion of participants in each treatment group with a recurrence at certain time points (6, 12, 24, 36, 48, and 72 months, where these data were available).
Risk factors associated with recurrent seizures, i.e. age at first seizure below 18 months, positive family history of seizures, temperature at index seizure below 40.0 °C.

Assessment of risk of bias in included studies

One review author (KV) made an initial assessment of all included studies for risk of bias using the Cochrane 'Risk of bias' tool for randomised or quasi‐randomised controlled trials (Higgins 2011). This was compared to an independent assessment by a second review author (RN or MO), with a third party resolving any disagreements by discussion.

Measures of treatment effect

We treated efficacy (recurrence of febrile or non‐febrile seizures) as dichotomous outcomes expressed as risk ratios (RR) with 95% confidence intervals (CIs).

We summarised treatment adherence and incidence of adverse effects narratively according to the definitions reported in the study. We calculated NNTBs as the reciprocal of the absolute risk reduction (McQuay 1998).

Unit of analysis issues

There were no unit of analysis issues. Medication dosages were standard. Outcome measures were simply seizure recurrence. No studies were of a repeated‐measure (longitudinal) nature or a cross‐over design.

Dealing with missing data

At times, recurrence data had to be reconstructed from published survival curves. We were careful to cross‐check these data with quoted cumulative incidence rates for in‐study data. We cross‐checked trial details against any additional published report of the trial and contacted the original trial authors in the case of missing data, errors, or inconsistencies (although the response was uniformly poor). No author provided individual‐participant data when requested; however, the consistency checks we performed were satisfactory.

Assessment of heterogeneity

We assessed clinical heterogeneity by reviewing the differences across trials in the characteristics of recruited participants and treatment protocols. We assessed statistical heterogeneity using a Chi² test for heterogeneity. We assessed heterogeneity using the Q test (P < 0.10 for significance) and the I² statistic (where greater than 50% indicated considerable heterogeneity) (Higgins 2003), and visually by inspecting forest plots.

Assessment of reporting biases

We assessed the presence of publication bias using funnel plots for each meta‐analysis that included results of five or more studies.

Data synthesis

We included studies comparing either different drugs or different treatment approaches, for example intermittent AED therapies versus continuous AED therapies, antipyretics during episodes of fever versus AED therapy during fever, or all versus placebo. The primary analysis was intention‐to‐treat and included all randomised participants analysed in the treatment groups to which they had been allocated, irrespective of which treatment they actually received.

We conducted meta‐analysis if sufficient data were available, that is at least two trials looking at the same two treatments and the same outcomes. All meta‐analyses were conducted using a fixed‐effect model, regardless of the presence of heterogeneity. If we had concerns regarding variability of study design and whether pooling data was appropriate, we would not have performed meta‐analysis.

We conducted meta‐analysis only for the primary outcome of efficacy (recurrence of febrile or non‐febrile seizures).

We summarised treatment adherence and incidence of adverse effects narratively according to the definitions reported in the study; we did not pool numerical data for these outcomes, due to variability in definitions and the level of detail reported in the studies.

Subgroup analysis and investigation of heterogeneity

We had no hypotheses needing subgroup analyses.

Sensitivity analysis

We considered there to be no need for any sensitivity analyses, as misdiagnosis of febrile seizures or their recurrence is unlikely within the reported study groups.

Summary of findings and assessment of the certainty of the evidence

We used the GRADE approach as outlined in the GRADE Handbook to interpret findings (Schünemann 2013), and GRADEpro GDT software (which imports data from Review Manager 5 software) to create 15 'Summary of findings' tables for the primary outcome of efficacy (recurrence of febrile or non‐febrile seizures) for each comparison for the following time points: 6 months, 12 months, 18 months, 24 months, 36 months, 48 months, and 60 or more months (GRADEpro GDT 2020; Review Manager 2020).

Results

Description of studies

Results of the search

See Figure 1.

Of 71 articles identified as potentially relevant, 32 articles met the inclusion criteria for the review (see Characteristics of included studies). Together, these 32 articles describe 32 randomised trials and their (long‐term) follow‐up. For the reasons for exclusion of the excluded studies, see Characteristics of excluded studies.

Included studies

The interventions compared against placebo or no treatment included intermittent oral diazepam in four studies (Autret 1990; Ramakrishnan 1986; Rosman 1993; Verrotti 2004) or rectal diazepam in five studies (Knudsen 1985; Mosquera 1987; Pavlidou 2006; Taghdiri 2011; Uhari 1995 (where a rectal dose was followed by oral doses for the time of the fever)), continuous phenobarbital in 10 studies (Bacon 1981; Camfield 1980; Farwell 1990; Heckmatt 1976; Mamelle 1984; McKinlay 1989; Ngwane 1980; Ramakrishnan 1986; Thilothammal 1993; Wolf 1977), intermittent phenobarbital in three studies (Mackintosh 1970; Ramakrishnan 1986; Wolf 1977), continuous oral phenytoin in one study (Bacon 1981), continuous oral valproate in five studies (Mamelle 1984; McKinlay 1989; Mosquera 1987; Ngwane 1980; Williams 1979), continuous oral pyridoxine in one study (McKiernan 1981), intermittent oral ibuprofen in one study (Van Stuijvenberg 1998), intermittent oral clobazam in one study (Bajaj 2005), continuous zinc sulfate for six months in one study (Fallah 2015), intermittent rectal diclofenac versus placebo followed by either ibuprofen or paracetamol or placebo after eight hours in one study (Strengell 2009), and intermittent oral levetiracetam versus any antipyretic given by way of a placebo (Hu 2014). Other studies compared interventions against each other: continuous phenobarbital versus intermittent diazepam in two studies (Garcia 1984; Salehiomran 2016); intermittent rectal diazepam versus intermittent rectal valproate in one study (Daugbjerg 1990); intermittent oral diazepam versus intermittent oral clobazam in two studies (Ghazavi 2016; Khosroshahi 2011); and intermittent oral melatonin versus intermittent oral diazepam in one study (Barghout 2019).

The included studies enrolled a total of 4542 participants with febrile seizures, of whom 4431 were used in the analysis of this review. The number of participants analysed for each intervention (number of participants included in placebo trials only) was as follows: diazepam 1476 (771); continuous phenobarbital 1075 (494); intermittent phenobarbital 341 (32); phenytoin 90 (90); valproate 303 (48); pyridoxine 107 (107); ibuprofen 230 (230); clobazam 60 (60); zinc sulfate 100 (100); diclofenac versus placebo followed by ibuprofen, paracetamol, or placebo after eight hours 231 (231); levetiracetam versus any unspecified antipyretic by way of a placebo 78 (37); continuous phenobarbital versus diazepam 245; diazepam versus valproate 169; diazepam versus clobazam 143; melatonin versus diazepam 60. It should be noted that a number of these papers included a comparison of outcomes in placebo versus one of two randomised seizure treatments (i.e. A versus C; B versus C). As no pooled analyses were done in which the effects of different antiepileptic or antipyretic drugs were summarised and compared with (placebo) controls, no unit of analysis errors were introduced. Families withdrew from these studies for various reasons, including change of residence, withdrawal of consent, and a variety of unacceptable adverse effects (detailed to the greatest degree possible in Table 16).

1. Treatment adherence.

Study	Treatment groups	Assessed	Method	Outcome	Treatment adjusted based on adherence assessment?
Autret 1990	‐DZP (oral) ‐PCB	Yes	Treatment diary	7% (1/15) of participants with relapses in DZP group were adherent versus 39% (7/18) in PCB group.	No
Bacon 1981	‐PT ‐PB (cont.) ‐PCB	Yes	Saliva and plasma	Recurrence was positively related to median drug levels for PB, but not related for PT. PB: 0/4 (0%) at < 5 mg/L; 5/19 (26%) at 5 to 8 mg/L; 5/25 (20%) at > 8 mg/L PT: 3/9 (33%) at < 0.5 mg/L, 9/19 (47%) at 0.5 to 1.0 mg/L, 4/19 (21%) at > 1.0 mg/L	Yes
Bajaj 2005	‐CBZ ‐PCB	No	‐	‐	‐
Barghout 2019	‐MEL ‐DZP (oral)	Yes	Return of used package of drugs during the follow‐up visits by parents	Not reported. There was 1 participant with non‐compliance in the primary sample (1/66), but this participant was not included in data analysis of the 60 participants.	No
Camfield 1980	‐PB (cont.) ‐PCB	Yes	Riboflavin urine check, and serum PB	Urine samples available in 65% (PB) and 56% (PCB); more than 90% of all samples tested positive. PB levels: mean 1.3 to 1.5 mg/dL, 70% to 81% within therapeutic range (≥ 1.0 mg/dL)	Yes
Daugbjerg 1990	‐DZP (rectal) ‐VP	No	‐	‐	‐
Fallah 2015	‐ZNC ‐PCB	No	‐	‐	‐
Farwell 1990	‐PB (cont.) ‐PCB	Yes	Riboflavin urine check, PB blood levels	Riboflavin results not reported. 2/3 (66%) of PB blood levels tested were above 645.9 μmol/L or 15 μg/mL.	Yes
Ghazavi 2016	‐CBZ ‐DZP (oral)	No	‐	‐	‐
Garcia 1984	‐DZP (rectal) ‐PB (cont.)	No	‐	‐	‐
Heckmatt 1976	‐PB (cont.) ‐NT	Yes	PB plasma levels	82% (40/49) had a mean PB plasma level above 65 μmol/L. All 4 recurrences in the PB group occurred in children with levels above 65 μmol/L.	Yes
Hu 2014	‐LEV (int.) ‐PCB	Yes	Parents/caregivers were contacted every 12 weeks.	Poor compliance in 4/78 (5.1%) LEV children and 1/37 (2.7%) PCB children	No
Khosroshahi 2011	‐DZP (oral) ‐CBZ	No	‐	‐	‐
Knudsen 1985	‐DZP (rectal) ‐NT	Yes	Historically in case of recurrence	Unclear report: "Parents treated the seizure as prescribed in 56/77 (72%) of the cases". Origin of the denominator is unclear, as 21 recurrences occurred in DZP and 77 in NT.	No
Mackintosh 1970	‐PB (int.) ‐PCB	No	‐	‐	‐
Mamelle 1984	‐PB (cont.) ‐VP ‐PCB	Yes	Blood levels	Unclear report	Yes
McKiernan 1981	‐PDX ‐PCB	Yes	Historically and counting of tablets used	Not reported	Yes
McKinlay 1989	‐PB (cont.) ‐VP ‐NT	Yes	PB and VP serum levels	PB: level checked 25/41 (61%) of children. Therapeutic level at time of recurrence: 5/12 (42%); level in those with non‐recurrence: 9/29 therapeutic, 11/29 subtherapeutic, 9/29 not done VP: level checked 36/50 (72%) of children. Therapeutic level at time of recurrence: 12/20 (60%); level in children with non‐recurrence: 13/30 therapeutic, 6/30 subtherapeutic, 11/30 not done	No
Mosquera 1987	‐DZP (rectal) ‐VP ‐NT	No	‐	‐
Ngwane 1980	‐PB (cont.) ‐VP	Yes	Blood levels (random moments)	35 measures in 28 of 39 included children (72%): 16 in PB, of which 4 (25%) below therapeutic range, and 19 in VP, of which 1 (5%) below therapeutic range	No
Pavlidou 2006	‐DZP (rectal) ‐NT	No	‐	‐	‐
Ramakrishnan 1986	‐PB (cont.) ‐PB (int.) ‐DZP (oral) ‐NT	No	‐	‐	‐
Rosman 1993	‐DZP (oral) ‐PCB	Yes	Riboflavin urine check	1257 DZP samples, 66% of all reported fever days, 96% of samples tested positive 982 PCB samples, 95% of all reported fever days, 95% of samples tested positive	No
Salehiomran 2016	‐DZP (oral) ‐PB (cont.)	No	‐	‐	‐
Strengell 2009	‐DCF ‐PCB	No	‐	‐	‐
Taghdiri 2011	‐DZP (rectal) ‐NT	No	‐	‐	‐
Thilothammal 1993	‐PB (cont.) ‐PCB	Yes	Counting sachets	"Poor compliance" in 2/30 (7%) PB children and in 1/30 (3%) PCB children. All children with "poor compliance" also had a recurrence.	No
Uhari 1995	‐DZP ‐PCB	No	‐	‐	‐
Van Stuijvenberg 1998	‐IBU ‐PCB	No	‐	‐	‐
Verrotti 2004	‐DZP (oral) ‐NT	Yes	Unclear, asked at recurrence	All 5 recurrences in DZP group were non‐compliant.	No
Williams 1979	‐VP ‐NT	Yes	Random VP plasma samples	Checked in 21/30 (70%) VP children: all showed measurable levels, but 2 below target concentration	No
Wolf 1977	‐PB (cont.) ‐PB (int.) ‐NT	Yes	4‐monthly blood check in the continuous PB group	78 of 106 cont. PB children (74%) had PB concentrations above target in at least 50% of their samples. These include 5 of the 7 children (71%) who had a recurrence in this group.	Yes

Intermittent oral or rectal diazepam compared to placebo or no treatment for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: intermittent oral or rectal diazepam Comparison: placebo or no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Outcomes	Risk with placebo or no treatment	Risk with intermittent oral or rectal diazepam	Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Recurrent seizure at 6 months	179 per 1000	115 per 1000 (86 to 152)	RR 0.64 (0.48 to 0.85)	1151 (6 RCTs)	⊕⊕⊕⊝ Moderate¹
Recurrent seizure at 12 months	254 per 1000	175 per 1000 (142 to 213)	RR 0.69 (0.56 to 0.84)	1416 (8 RCTs)	⊕⊕⊕⊝ Moderate¹
Recurrent seizure at 18 months	336 per 1000	124 per 1000 (77 to 201)	RR 0.37 (0.23 to 0.60)	289 (1 RCT)	⊕⊕⊝⊝ Low²
Recurrent seizure at 24 months	273 per 1000	200 per 1000 (153 to 260)	RR 0.73 (0.56 to 0.95)	739 (4 RCTs)	⊕⊕⊕⊕ High
Recurrent seizure at 36 months	606 per 1000	351 per 1000 (242 to 515)	RR 0.58 (0.40 to 0.85)	139 (1 RCT)	⊕⊕⊝⊝ Low²
Recurrent seizure at 48 months	308 per 1000	111 per 1000 (46 to 274)	RR 0.36 (0.15 to 0.89)	110 (1 RCT)	⊕⊕⊕⊝ Moderate³
Recurrent seizure at 60 months or greater	200 per 1000	16 per 1000 (0 to 262)	RR 0.08 (0.00 to 1.31)	60 (1 RCT)	⊕⊝⊝⊝ Very low^2,4
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Continuous phenobarbital compared to placebo or no treatment for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: continuous phenobarbital Comparison: placebo or no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Outcomes	Risk with placebo or no treatment	Risk with continuous phenobarbital	Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Recurrent seizure at 6 months	178 per 1000	105 per 1000 (75 to 148)	RR 0.59 (0.42 to 0.83)	833 (6 RCTs)	⊕⊕⊕⊝ Moderate¹
Recurrent seizure at 12 months	308 per 1000	166 per 1000 (129 to 216)	RR 0.54 (0.42 to 0.70)	807 (7 RCTs)	⊕⊕⊝⊝ Low^1,2
Recurrent seizure at 18 months	430 per 1000	331 per 1000 (241 to 451)	RR 0.77 (0.56 to 1.05)	264 (2 RCTs)	⊕⊕⊕⊝ Moderate¹
Recurrent seizure at 24 months	345 per 1000	238 per 1000 (183 to 307)	RR 0.69 (0.53 to 0.89)	533 (3 RCTs)	⊕⊕⊕⊝ Moderate¹
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	200 per 1000	300 per 1000 (122 to 738)	RR 1.50 (0.61 to 3.69)	60 (1 RCT)	⊕⊝⊝⊝ Very low^3,4
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Intermittent phenobarbital compared to placebo or no treatment for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: intermittent phenobarbital Comparison: placebo or no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Outcomes	Risk with placebo or no treatment	Risk with intermittent phenobarbital	Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Recurrent seizure at 6 months	88 per 1000	121 per 1000 (59 to 247)	RR 1.37 (0.67 to 2.81)	281 (2 RCTs)	⊕⊝⊝⊝ Very low^1,2,3
Recurrent seizure at 12 months	216 per 1000	218 per 1000 (140 to 343)	RR 1.01 (0.65 to 1.59)	281 (2 RCTs)	⊕⊕⊕⊝ Moderate¹
Recurrent seizure at 18 months	Not reported				NA
Recurrent seizure at 24 months	294 per 1000	250 per 1000 (167 to 376)	RR 0.85 (0.57 to 1.28)	249 (1 RCT)	⊕⊕⊝⊝ Low⁴
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	200 per 1000	166 per 1000 (56 to 488)	RR 0.83 (0.28 to 2.44)	60 (1 RCT)	⊕⊝⊝⊝ Very low^3,4
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Continuous oral phenytoin compared to placebo for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: continuous oral phenytoin Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with placebo	Risk with continuous oral phenytoin
Recurrent seizure at 6 months	Not reported				NA
Recurrent seizure at 12 months	349 per 1000	342 per 1000 (192 to 603)	RR 0.98 (0.55 to 1.73)	90 (1 RCT)	⊕⊕⊝⊝ Low¹
Recurrent seizure at 18 months	Not reported				NA
Recurrent seizure at 24 months	Not reported				NA
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Continuous oral valproate compared to placebo or no treatment for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: continuous oral valproate Comparison: placebo or no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Outcomes	Risk with placebo or no treatment	Risk with continuous oral valproate	Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Recurrent seizure at 6 months	118 per 1000	141 per 1000 (65 to 308)	RR 1.20 (0.55 to 2.62)	156 (2 RCTs)	⊕⊕⊝⊝ Low¹
Recurrent seizure at 12 months	239 per 1000	196 per 1000 (124 to 308)	RR 0.82 (0.52 to 1.29)	255 (4 RCTs)	⊕⊕⊝⊝ Low¹
Recurrent seizure at 18 months	346 per 1000	45 per 1000 (7 to 332)	RR 0.13 (0.02 to 0.96)	48 (1 RCT)	⊕⊝⊝⊝ Very low^1,2
Recurrent seizure at 24 months	212 per 1000	267 per 1000 (155 to 462)	RR 1.26 (0.73 to 2.18)	156 (2 RCTs)	⊕⊕⊝⊝ Low¹
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Continuous oral pyridoxine compared to placebo for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: continuous oral pyridoxine Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with placebo	Risk with continuous oral pyridoxine
Recurrent seizure at 6 months	154 per 1000	72 per 1000 (23 to 228)	RR 0.47 (0.15 to 1.48)	107 (1 RCT)	⊕⊕⊝⊝ Low^1,2
Recurrent seizure at 12 months	192 per 1000	127 per 1000 (52 to 310)	RR 0.66 (0.27 to 1.61)	107 (1 RCT)	⊕⊕⊝⊝ Low^1,2
Recurrent seizure at 18 months	Not reported				NA
Recurrent seizure at 24 months	Not reported				NA
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Intermittent oral ibuprofen compared to placebo for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: intermittent oral ibuprofen Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with placebo	Risk with intermittent oral ibuprofen
Recurrent seizure at 6 months	210 per 1000	233 per 1000 (145 to 380)	RR 1.11 (0.69 to 1.81)	230 (1 RCT)	⊕⊕⊕⊕ High
Recurrent seizure at 12 months	294 per 1000	279 per 1000 (185 to 421)	RR 0.95 (0.63 to 1.43)	230 (1 RCT)	⊕⊕⊕⊕ High
Recurrent seizure at 18 months	Not reported				NA
Recurrent seizure at 24 months	387 per 1000	325 per 1000 (228 to 460)	RR 0.84 (0.59 to 1.19)	230 (1 RCT)	⊕⊕⊕⊕ High
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Intermittent oral clobazam compared to placebo for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: intermittent oral clobazam Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with placebo	Risk with intermittent oral clobazam
Recurrent seizure at 6 months	833 per 1000	300 per 1000 (167 to 533)	RR 0.36 (0.20 to 0.64)	60 (1 RCT)	⊕⊕⊝⊝ Low^1,2
Recurrent seizure at 12 months	Not reported				NA
Recurrent seizure at 18 months	Not reported				NA
Recurrent seizure at 24 months	Not reported				NA
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Continuous zinc sulfate for 6 months compared to placebo for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: continuous zinc sulfate for 6 months Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with placebo	Risk with continuous zinc sulfate for 6 months
Recurrent seizure at 6 months	Not reported				NA
Recurrent seizure at 12 months	380 per 1000	220 per 1000 (118 to 414)	RR 0.58 (0.31 to 1.09)	100 (1 RCT)	⊕⊕⊕⊕ High
Recurrent seizure at 18 months	Not reported				NA
Recurrent seizure at 24 months	Not reported				NA
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Intermittent rectal diclofenac compared to placebo, followed by oral ibuprofen, paracetamol, or placebo after 8 hours, for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: intermittent rectal diclofenac followed by oral ibuprofen, paracetamol, or placebo after 8 hours Comparison: placebo followed by oral ibuprofen, paracetamol, or placebo after 8 hours
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Outcomes	Risk with placebo followed by oral ibuprofen, paracetamol, or placebo after 8 hours	Risk with intermittent rectal diclofenac followed by oral ibuprofen, paracetamol, or placebo after 8 hours	Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
Recurrent seizure at 6 months	149 per 1000	119 per 1000 (63 to 231)	RR 0.80 (0.42 to 1.55)	231 (1 RCT)	⊕⊕⊕⊕ High
Recurrent seizure at 12 months	237 per 1000	163 per 1000 (95 to 275)	RR 0.69 (0.40 to 1.16)	231 (1 RCT)	⊕⊕⊕⊕ High
Recurrent seizure at 18 months	272 per 1000	196 per 1000 (122 to 315)	RR 0.72 (0.45 to 1.16)	231 (1 RCT)	⊕⊕⊕⊕ High
Recurrent seizure at 24 months	281 per 1000	222 per 1000 (143 to 348)	RR 0.79 (0.51 to 1.24)	231 (1 RCT)	⊕⊕⊕⊕ High
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Recurrent seizure @ 6 months	6	1151	Risk Ratio (M‐H, Fixed, 95% CI)	0.64 [0.48, 0.85]
1.1.1 Intermittent oral diazepam	2	516	Risk Ratio (M‐H, Fixed, 95% CI)	0.70 [0.45, 1.11]
1.1.2 Intermittent rectal diazepam	4	635	Risk Ratio (M‐H, Fixed, 95% CI)	0.59 [0.41, 0.86]
1.2 Recurrent seizure @ 12 months	8	1416	Risk Ratio (M‐H, Fixed, 95% CI)	0.69 [0.56, 0.84]
1.2.1 Intermittent oral diazepam	3	701	Risk Ratio (M‐H, Fixed, 95% CI)	0.72 [0.53, 0.99]
1.2.2 Intermittent rectal diazepam	5	715	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.50, 0.86]
1.3 Recurrent seizure @ 18 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.3.1 Intermittent rectal diazepam	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.4 Recurrent seizure @ 24 months	4	739	Risk Ratio (M‐H, Fixed, 95% CI)	0.73 [0.56, 0.95]
1.4.1 Intermittent oral diazepam	2	516	Risk Ratio (M‐H, Fixed, 95% CI)	0.62 [0.45, 0.85]
1.4.2 Intermittent rectal diazepam	2	223	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.67, 1.90]
1.5 Recurrent seizure @ 36 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.5.1 Intermittent rectal diazepam	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.6 Recurrent seizure @ 48 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.6.1 Intermittent oral diazepam	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.7 Recurrent seizure @ 60 to 72 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.7.1 Intermittent oral diazepam	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Recurrent seizure @ 6 months	6	833	Risk Ratio (M‐H, Fixed, 95% CI)	0.59 [0.42, 0.83]
2.2 Recurrent seizure @ 12 months	7	807	Risk Ratio (M‐H, Fixed, 95% CI)	0.54 [0.42, 0.70]
2.3 Recurrent seizure @ 18 months	2	264	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.56, 1.05]
2.4 Recurrent seizure @ 24 months	3	533	Risk Ratio (M‐H, Fixed, 95% CI)	0.69 [0.53, 0.89]
2.5 Recurrent seizure @ 60 to 72 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
2.6 Behavioural changes	1	65	Risk Ratio (M‐H, Fixed, 95% CI)	1.61 [0.79, 3.26]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Recurrent seizure @ 6 months	2	281	Risk Ratio (M‐H, Fixed, 95% CI)	1.37 [0.67, 2.81]
3.2 Recurrent seizure @ 12 months	2	281	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.65, 1.59]
3.3 Recurrent seizure @ 24 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
3.4 Recurrent seizure @ 60 to 72 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Recurrent seizure @ 6 months	2	156	Risk Ratio (M‐H, Fixed, 95% CI)	1.20 [0.55, 2.62]
5.2 Recurrent seizure @ 12 months	4	255	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.52, 1.29]
5.3 Recurrent seizure @ 18 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
5.4 Recurrent seizure @ 24 months	2	156	Risk Ratio (M‐H, Fixed, 95% CI)	1.26 [0.73, 2.18]

Continuous phenobarbital compared to intermittent rectal or oral diazepam for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: continuous phenobarbital Comparison: intermittent rectal or oral diazepam
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with intermittent rectal or oral diazepam	Risk with continuous phenobarbital
Recurrent seizure at 6 months	Not reported				NA
Recurrent seizure at 12 months	155 per 1000	229 per 1000 (116 to 455)	RR 1.48 (0.75 to 2.94)	145 (1 RCT)	⊕⊕⊝⊝ Low¹
Recurrent seizure at 18 months	80 per 1000	100 per 1000 (29 to 350)	RR 1.25 (0.36 to 4.38)	100 (1 RCT)	⊕⊝⊝⊝ Very low^1,2
Recurrent seizure at 24 months	Not reported				NA
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Intermittent rectal diazepam compared to intermittent rectal valproate for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: intermittent rectal diazepam Comparison: intermittent rectal valproate
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with intermittent rectal valproate	Risk with intermittent rectal diazepam
Recurrent seizure at 6 months	88 per 1000	123 per 1000 (51 to 304)	RR 1.41 (0.58 to 3.47)	169 (1 RCT)	⊕⊝⊝⊝ Very low^1,2
Recurrent seizure at 12 months	175 per 1000	259 per 1000 (144 to 467)	RR 1.48 (0.82 to 2.67)	169 (1 RCT)	⊕⊕⊝⊝ Low¹
Recurrent seizure at 18 months	Not reported				NA
Recurrent seizure at 24 months	Not reported				NA
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Intermittent oral diazepam compared to oral clobazam for febrile seizures in children
Patient or population: children with febrile seizures Setting: outpatients Intervention: intermittent oral diazepam Comparison: oral clobazam
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with intermittent oral clobazam	Risk with intermittent oral diazepam
Recurrent seizure at 6 months	Not reported				NA
Recurrent seizure at 12 months	42 per 1000	96 per 1000 (26 to 356)	RR 2.28 (0.62 to 8.42)	143 (2 RCTs)	⊕⊕⊝⊝ Low^1,2
Recurrent seizure at 18 months	Not reported				NA
Recurrent seizure at 24 months	Not reported				NA
Recurrent seizure at 36 months	Not reported				NA
Recurrent seizure at 48 months	Not reported				NA
Recurrent seizure at 60 months or greater	Not reported				NA
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk (the event rate in the control group) and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Intermittent oral melatonin compared with intermittent oral diazepam for febrile seizures in children
Patient or population: children with febrile seizures Settings: outpatients Intervention: intermittent oral melatonin Comparison: intermittent oral diazepam
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with oral diazepam	Risk with oral melatonin
Recurrent seizures at 6 months	367 per 1000	167 per 1000 (66 to 422)	RR 0.45 (0.18 to 1.15)	60 (1 RCT)	⊕⊝⊝⊝ Very low^1,2
Recurrent seizures at 12 months	Not reported				NA
Recurrent seizures at 18 months	Not reported				NA
Recurrent seizures at 24 months	Not reported				NA
Recurrent seizures at 36 months	Not reported				NA
Recurrent seizures at 48 months	Not reported				NA
Recurrent seizures at 60 months or greater	Not reported				NA
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Intermittent oral levetiracetam compared with placebo for febrile seizures in children
Patient or population: children with febrile seizures Settings: outpatients Intervention: intermittent oral levetiracetam Comparison: placebo (any antipyretics)
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)
	Risk with placebo	Risk with oral levetiracetam
Recurrent seizures at 6 months	Not reported				NA
Recurrent seizures at 12 months	514 per 1000	141 per 1000 (77 to 267)	RR 0.27 (0.15 to 0.52)	115 (1 RCT)	⊕⊝⊝⊝ Very low¹
Recurrent seizures at 18 months	Not reported				NA
Recurrent seizures at 24 months	Not reported				NA
Recurrent seizures at 36 months	Not reported				NA
Recurrent seizures at 48 months	Not reported				NA
Recurrent seizures at 60 months or greater	Not reported				NA
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

First author	Number of children	Adverse medication effects, as reported in article
Autret 1990	177	Hyperactivity (defined as agitation and inability to remain still), significantly (P < 0.003) more frequent in diazepam group (138 vs 34 days). No significant differences noted for normal vigilance or drowsiness; normal staggering or impossible "walking". 1 sudden unexpected death in placebo group
Bacon 1981	138, 43 control, 48 phenobarbital, 47 phenytoin	Rash in 1 child on phenobarbital, ataxia in 5 children on phenytoin. Behavioural items: whinginess; crying a lot, bad temper, tantrums, dislike of being left, unsteadiness, desire for cuddling, difficulty feeding, noisiness, thumb sucking. No significant difference for any of these items between phenobarbital/phenytoin or placebo group. Any behavioural change attributed to hospitalisation.
Bajaj 2005	60	Drug reactions Group A (clobazam) Group B (placebo); n (%) n (%): weakness 1 (3.3) 11 (33.3); irritability 4 (13.3) 1 (3.3); sedation 5 (16.7) 5 (16.7); anorexia 2 (6.6) 5 (16.7); nausea and vomiting 0, 2 (6.6); abdominal pain 0, 1 (3.3); diarrhoea 1 (3.3) 3 (10); headache 1 (3.3) 5 (16.7)
Barghout 2019	60	Oral diazepam 7/30 (23.3%): somnolence 3 (10%); ataxia/dizziness 2 (6.7%); vomiting 1 (3.3%); irritability 1 (3.3%) Oral melatonin 4/30 (13.3%): vomiting 3 (10%); diarrhoea 1 (3.3%)
Camfield 1980	79	At 12 months there was no difference between phenobarbital and placebo groups for behavioural change or sleep disturbance. Placebo group: transient adverse effects in 8 of 30; phenobarbital group: transient adverse effects 15 of 35. Significant negative correlation between phenobarbital serum level and memory concentration subscores on Binet scores. Lower comprehension scores showed significant correlation with length of phenobarbital treatment (but n = 7 at 8 months and n = 9 at 12 months, therefore small numbers).
Daugbjerg 1990	169	Diazepam seen in 42 (47%), as follows: sedation 33 (37%), ataxia 42 (47%), hyperkinesia 21 (24%), diarrhoea, urge to defecate 1 (1%), depression 1 (1%). Valproic acid: sedation 9 (11%), ataxia 3 (4%), hyperkinesia 6 (7%), diarrhoea, urge to defecate 14 (18%). Vomiting 1 (1%), bleeding per rectum 1 (1%), abdominal pain 3 (4%), aggressiveness 3 (4%)
Fallah 2015	100	No serious side effects were witnessed in the 2 groups. Gastrointestinal side effects including vomiting in 5 (10%) children, heartburn in 2 (4%), and abdominal pain in 1 (2%) child were seen in 16% of the zinc sulfate group. All of the side effects were well tolerated and disappeared in 2 to 3 weeks, and supplementation continued. Vomiting occurred in 2 children (4%) in the control group.
Farwell 1990	217	Investigators compared intelligence quotients (IQs) of a group randomly assigned to phenobarbital to a group randomly assigned to placebo. After 2 years, mean IQ 8.4 points lower in phenobarbital group (95% CI ‐13.3 to −3.5, P = 0.006). 6 months later after discontinuing medication, IQ 5.2 points lower in phenobarbital group (95% CI −10.5 to 0.04, P = 0.052). Proportion remaining seizure‐free did not differ significantly between treatment groups. 14 total sleep time, night awakenings, and lengthy awakenings compared between phenobarbital and placebo groups; no difference noted between groups, except subset of predisposed children did experience an increase in night awakenings (i.e. those already recorded to have frequent sleep disturbances at study entry). 35: retesting of group after school entry. Phenobarbital‐treated group had Wide Range Achievement Test (WRAT‐R) reading achievement score significantly lower than placebo group: 87.6 vs 95.6; P = 0.007. No significant difference for IQ on Stamford Binet
Garcia 1984	100	Adverse effects: diazepam 5 (10%), phenobarbital 3 (6%). Nature of adverse effects not stated.
Ghazavi 2016	71	Ataxia: diazepam 4/35 (11%), clobazam 1/36 (3%)
Heckmatt 1976	161	Overall, 39 of 88 stopped taking phenobarbital: 16 behaviour (overactivity, unpleasant behaviour, temper, not sleeping) (12 improved); 23 for a variety of reasons, e.g. drowsy/unsteady. 3 in control group reported behaviour problems.
Hu 2014	115	Levetiracetam group: 1/78 (1.2%) somnolence; control group: no adverse effects
Knudsen 1985	152	No severe adverse effects. Mild transient: 36% sedation, 15% euphoria, 8% ataxia, 2% aggression. Adverse effects not addressed in report on follow‐up.
Khosroshahi 2011	72	Adverse effects of clobazam were noted to be lower than with diazepam. Sedation was noted more often with diazepam compared to clobazam (P < 0.001). No further details given.
Mamelle 1984		Adverse effects not addressed.
Mackintosh 1970	32	Adverse effects not addressed.
McKiernan 1981	107	Adverse effects not addressed.
McKinlay 1989	151	13 of 41 on phenobarbital had disturbed behaviour or drowsiness, or both; 1 vomiting; 2 rash; 1 unacceptable taste. 8 stopped treatment, 3 within 3 months. 5 of 50 on valproate drowsy initially; 2 behavioural problems; 1 vomited; 1 diarrhoea. 2 stopped taking drug. 16 control group adverse effects not addressed.
Mosquera 1987	69	Adverse effects not addressed.
Ngwane 1980	43	5 of 23 on phenobarbital had adverse effects within 72 hours; drug was withdrawn in 2 of these (details not given). 4 of 20 on valproate had adverse effects, most commonly diarrhoea.
Pavlidou 2006	139	Adverse effects were only reported in the diazepam group. These were described as mild and transient and included somnolence and irritability.
Ramakrishnan 1986	120	Adverse effects not addressed.
Rosman 1993	288	Of 135 children on placebo, 1 “moderate” maculopapular rash. Of 153 on diazepam, 59 (39%) at least moderate adverse effects: ataxia 30%, lethargy 29%, irritability 24%; moderate adverse effects: unclear speech 6%, hyperactivity 6%, insomnia 5%, hallucinations 0.7% (percentages of those 59 (39%) overall who had adverse effects). Mild adverse effects paralleled moderate numbers.
Salehiomran 2016	145	Side effects of phenobarbital like hyperkinesia, irritability, and restlessness were observed in some children, but diazepam‐related side effects, except for sedation, were not seen.
Strengell 2009	231	Adverse effects not addressed.
Van Stuijvenberg 1998	230	Adverse effects not addressed.
Thilothammal 1993	90	“Intolerable” side effects presented in 2 of 30 children with simple febrile seizures on phenobarbital and 1 of 30 children with atypical febrile seizures. Recorded adverse effects were “mainly hyperkinetic behaviour, extreme irritability, fussiness and aggressiveness”. Details of percentages are not given.
Uhari 1995	180	Adverse effects not addressed.
Verrotti 2004	110	Adverse effects were only reported in the treatment group, and included ataxia, lethargy, and irritability: 14 children (31.1%) had ataxia, 13 (28.8%) presented lethargy, and 11 children (24.4%) had irritability. These adverse effects lasted no more than 36 hours.
Williams 1979	58	7 of 30 children taking valproate (23%) had adverse effects: 4 diarrhoea or vomiting; 1 increased appetite; 1 increased daytime activity, night terrors, and confusion; 1 anorexia, withdrawn and crying. Adverse effects in control group not detailed.
Wolf 1977	355	Phenobarbital 34 of 109 (32%) discontinued continuous phenobarbital, reasons as follows: 16% hyperactivity; 1% irritability; 3% rash; 2% lethargy; 10% parental non‐compliance. Long‐term effect of phenobarbital on cognitive function: group of 50 matched for age, sex, rash, and socio‐economic status for difference in cognitive function to median age of 57.5 months (phenobarbital‐treated children) and 59.6 months (children not receiving phenobarbital).

Date	Event	Description
3 February 2020	New citation required but conclusions have not changed	Conclusions are unchanged.
3 February 2020	New search has been performed	Searches updated 3 February 2020; two new studies identified and included.

Date	Event	Description
21 July 2016	New citation required but conclusions have not changed	Conclusions are unchanged.
21 July 2016	New search has been performed	Searches updated 21 July 2016; four new studies were identified and added as included studies in the review.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Recurrent seizure @ 6 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
6.2 Recurrent seizure @ 12 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	Statistical method	Effect size
7.1 Recurrent seizure @ 6 months	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
7.2 Recurrent seizure @ 12 months	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
7.3 Recurrent seizure @ 24 months	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

PERMALINK

Prophylactic drug management for febrile seizures in children

Martin Offringa

Richard Newton

Sarah J Nevitt

Katerina Vraka

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Methodological and trial design

Participant and demographic information

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Summary of findings and assessment of the certainty of the evidence

Results

Description of studies

Results of the search

1.

Included studies

1. Treatment adherence.

A brief description of the 32 original studies reported in the articles included in this review

Excluded studies

Risk of bias in included studies

2.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Publication bias

1.1. Analysis.

1.2. Analysis.

2.1. Analysis.

2.2. Analysis.

3.

4.

5.

6.

5.2. Analysis.

Effects of interventions

Summary of findings 1. Intermittent oral or rectal diazepam compared to placebo or no treatment for febrile seizures in children.

Summary of findings 2. Continuous phenobarbital compared to placebo or no treatment for febrile seizures in children.

Summary of findings 3. Intermittent phenobarbital compared to placebo or no treatment for febrile seizures in children.

Summary of findings 4. Continuous oral phenytoin compared to placebo for febrile seizures in children.

Summary of findings 5. Continuous oral valproate compared to placebo or no treatment for febrile seizures in children.

Summary of findings 6. Continuous oral pyridoxine compared to placebo for febrile seizures in children.

Outcome or subgroup title	No. of studies	Statistical method	Effect size
10.1 Recurrent seizure @ 6 months	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
10.2 Recurrent seizure @ 12 months	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
10.3 Recurrent seizure @ 18 months	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
10.4 Recurrent seizure @ 24 months	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
11.1 Recurrent seizure @ 12 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
11.2 Recurrent seizure @ 18 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
12.1 Recurrent seizure @ 6 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
12.2 Recurrent seizure @ 12 months	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

*Study characteristics*
Methods	Double‐blind RCT
Participants	185, age 8 to 36 months, first FS, < 2 RF
Interventions	Intermittent oral diazepam, 0.5 mg load, 0.2 mg maintenance per kilo, or placebo
Outcomes	RS at 12 months, adverse effects at 12 months
Notes	Attrition: 6 diazepam, 3 placebo; results presented as participant days; significant hyperactivity in diazepam group; 1 SUDEP in placebo group
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	Centralised allocation
Blinding (performance bias and detection bias) All outcomes	Low risk	Double‐blind
Incomplete outcome data (attrition bias) All outcomes	Low risk	9 (6 diazepam, 3 placebo) of 185 withdrawn
Selective reporting (reporting bias)	Low risk	Stated outcome objective met.
Other bias	Low risk	No bias identified.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Double‐blind

*Study characteristics*
Methods	RCT, open‐label
Participants	169, first FS
Interventions	Rectal diazepam 5 mg for < 3 years; 7.5 mg for 3 years or over; or valproate suppository 150 mg for < 10 kg; 300 mg for 10 kg or more
Outcomes	RS at 6 months, 12 months, adverse effects
Notes	2 withdrawn, 4 lost during follow‐up
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	High risk	Odd/even dates ‐ no concealment
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding (selection bias)
Incomplete outcome data (attrition bias) All outcomes	Low risk	6 of 169 withdrawn; 4 lost to follow‐up in each group
Selective reporting (reporting bias)	Low risk	Stated outcome objective met.
Other bias	Low risk	No bias identified.
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding
Blinding of outcome assessment (detection bias) All outcomes	High risk	No blinding

*Study characteristics*
Methods	Single‐centre, randomised, single‐blind clinical study
Participants	Children aged 1.5 to 5 years with first simple FS, weight and height above the third percentile, and normal serum zinc level
Interventions	Group 1: Daily zinc sulfate 2 mg/kg (maximum 50 mg) for 6 consecutive months Group 2: Placebo
Outcomes	Seizure recurrence at 12 months, side effects
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	Computer‐generated equal simple randomisation
Blinding (performance bias and detection bias) All outcomes	Low risk	Single‐blind design
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up, no exclusions
Selective reporting (reporting bias)	High risk	Recurrence data at 3, 6, and 9 months not given. Kaplan‐Meier method used to report results, no absolute numbers.
Other bias	Low risk	No bias identified.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Randomisation and blinding were done by an investigator with no clinical involvement in the trial. Data collectors, outcome assessors, and data analysts were all kept blinded to the allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Randomisation and blinding were done by an investigator with no clinical involvement in the trial. Data collectors, outcome assessors, and data analysts were all kept blinded to the allocation.

*Study characteristics*
Methods	Single‐centre, randomised, open‐label trial
Participants	Children 6 to 60 months of age with at least 1 simple FS
Interventions	Oral diazepam 0.33 mg/kg every 8 hours for 2 days or oral clobazam for 2 days dosed by child's weight (daily 5 mg when weight ≤ 5 kg, twice‐daily 5 mg when 6 to 10 kg, twice‐daily 7.5 mg when 11 to 15 kg, twice‐daily 10 mg when > 15 kg)
Outcomes	RS at 12 months and adverse effects
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	High risk	Randomisation methodology not mentioned.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding
Incomplete outcome data (attrition bias) All outcomes	High risk	Not discussed
Selective reporting (reporting bias)	Unclear risk	Unclear
Other bias	Low risk	No bias identified.
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding
Blinding of outcome assessment (detection bias) All outcomes	High risk	No blinding

*Study characteristics*
Methods	Quasi‐RCT
Participants	165, first FS, mean age 20 months
Interventions	Phenobarbital 4 to 5 mg per kilo or no treatment
Outcomes	RS at 6 months
Notes	Attrition: 4 (2 per arm), unblinded study
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	High risk	Alternate‐day allocation
Blinding (performance bias and detection bias) All outcomes	High risk	None
Incomplete outcome data (attrition bias) All outcomes	Low risk	4 of 165 lost, but 39 of 88 stopped treatment
Selective reporting (reporting bias)	Low risk	Stated outcome objective met.
Other bias	Low risk	No bias identified.
Blinding of participants and personnel (performance bias) All outcomes	High risk	None
Blinding of outcome assessment (detection bias) All outcomes	High risk	None

*Study characteristics*
Methods	Unblinded RCT, block randomisation
Participants	115 children, with onset age between 3 months and 5 years, with a history of 2 or more episodes of FS within the last 6 months
Interventions	Treatment group: oral levetiracetam first week 15 to 30 mg/kg/d in 2 divided doses; second week: dose tapering 50% every 2 days Control group: antipyretic when temperature > 38.5 °C with or without antibiotics
Outcomes	Primary outcome: seizure frequency associated with febrile events and FS recurrence rate during the 48‐week follow‐up Second outcome: cost‐effectiveness of the 2 groups
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	Computer‐generated block randomisation Managed by centralised control who managed the randomisation code Sequentially numbered envelope corresponding to the individual
Blinding (performance bias and detection bias) All outcomes	High risk	Unblinded (both physicians and parents) Unblinded outcome assessors
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	115 randomised (2:1) and 22 excluded (19.1% loss); increased losses 18 in the levetiracetam group 4 in the control group No significant differences between 2 groups in the constituent ratio
Selective reporting (reporting bias)	Low risk	Clearly defined primary outcome with clear reporting
Other bias	Unclear risk	Did not divide levetiracetam group into high and low dose but used a range of doses Suggested to use any antipyretic (not active control ‐ pseudo placebo)
Blinding of participants and personnel (performance bias) All outcomes	High risk	Both were unblinded.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Unblinded outcome assessors

*Study characteristics*
Methods	Single‐blind RCT
Participants	69, 6 to 48 months, first FS, excluded focal and neuropsychiatric disorders
Interventions	Phenobarbital 3 to 4 mg per kilo, or valproate 30 to 40 mg per kilo, or placebo
Outcomes	RS at 18 months, length of follow‐up differed (mean > 20 months)
Notes	Attrition: 4
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	High risk	Not used
Blinding (performance bias and detection bias) All outcomes	High risk	Unblinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	4 of 69 dropped out.
Selective reporting (reporting bias)	Low risk	Stated outcome objective met.
Other bias	Low risk	No bias identified.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Unblinded