Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Does oral diazepam prevent febrile convulsions in children?

Three Part Question

In [children with a previous febrile seizure] does [regular oral diazepam]reduce [recurrence of febrile seizures]?

Clinical Scenario

An 18-month-old boy is admitted with his first febrile convulsion. At discharge his parents ask if there is there anything that can be done to prevent another one?

Search Strategy

Primary resources: Pubmed. (1966 – Aug 2003):
Fever AND seizures AND prevention, Midazolam AND seizure, febrile Diazepam AND seizure, febrile.
Secondary Sources: Cochrane library and DARE – ''fever AND seizures AND prevention'', ''midazolam AND seizure, febrile'', ''diazepam AND seizure, febrile''.
Prodigy Evidence Based Clinical Guidelines – 'febrile convulsions'

Search Outcome

Pubmed. (1966 – Aug 2003):
Fever AND seizures AND prevention 148
Midazolam AND seizure, febrile 8
Diazepam AND seizure, febrile 125

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Rosman R et al
406 children with previous febrile seizure a) Oral diazepam 0.3mg/Kg 8hrly b) PlaceboRandomised double blind placebo controlled trial (level 1 b)Recurrence of febrile Seizures.a - 41(20%). b - 72(35%) Chi Squared 10.6, P 0.001 RRR 0.425 (CI 0.002, 0.587) ARR 0.915 (CI 0.064,0.236), NNT 7 (nnh=4, nnt=16)Good study but diazepam was not easy to use, and lots of side effects were reported.
185 children with a first febrile seizure a) Oral diazepam 0.5mg/Kg then 0.2mg/Kg 12 hourly b) PlaceboRandomised double blind placebo controlled trial (level 1 b)Recurrence of febrile Seizuresa - 15(16%). b - 18(19.5%) Chi squared 0.175 P value 0.676 RRR 0.176 (CI -0.535, 0.557) ARR 0.034 (CI -0.076, 0.145) NNT 29 (nnh=7, nnt=13)Hyperactivity was significantly more frequent in the diazepam group (p value 0.003). Follow up was relatively short at 10.3 months.


Rosman et al. (1993) randomised 406 children with febrile seizures, to receive oral diazepam or placebo regularly during subsequent febrile illnesses. Febrile seizures occurred in 41(20%) of those receiving diazepam and 73(35%) of the control group. Whilst the total number of seizures in the diazepam compared to the control group when taking the medication was 7 Vs 38, the number needed to treat (NNT) to prevent further seizures was 7 (see table). Amongst 204 receiving placebo 1(0.4%) had moderate side effects, whereas 59(29%) in the diazepam arm experienced ataxia, lethargy and irritability and 5% experienced changes in speech, activity level and sleep. Furthermore, compliance with medication was poor with 83%(34) seizures occurring whilst not receiving the diazepam schedule. The reasons for this were that parents did not measure the temperature, convulsions were the first sign of illness or the last temperature was normal. Autret at al. (1990) randomised children with previous febrile seizures to oral diazepam or placebo, with a mean duration of follow up of 10-months: 16% in the diazepam arm and 19.5% receiving placebo had febrile seizures, the NNT was 29 (see table) and hyperactivity was significantly more frequent in the diazepam group (p value 0.003).

Clinical Bottom Line

Oral diazepam it is not suitable for prophylaxis in all children with febrile seizures because of difficulties with compliance, frequent adverse effects and a relatively large NNT.


  1. Rosman NP, Colton T, Labazzo J et al. A controlled trial of diazepam administered during febrile illnesses to prevent recurrence of febrile seizures. NEJM 1993;329(2):79-84.
  2. Autret E, Billard C, Motte J et al. Double-blind, randomised trial of diazepam versus placebo for prevention of recurrence of febrile seizures. J Pediatr 1990;117(3):490-4.