Intermittent diazepam and continuous phenobarbital to treat recurrence of febrile seizures: a systematic review with meta-analysis

Diazepam intermitente e fenobarbital contínuo para tratamento da recorrência de convulsões febris: uma revisão sistemática com metanálise

Intermittent diazepam and continuous phenobarbital to treat recurrence of febrile seizures: A systematic review with meta-analysis

Diazepam intermitente e fenobarbital contínuo para tratamento da recorrência de convulsões febris: uma revisão sistemática com metanálise

Alice Hatsue Masuko^I; Aldemar Araujo Castro^II; Gustavo Rocha Santos^III; Álvaro Nagib Atallah^III; Lucila Bizari Fernandes do Prado^IV; Luciane Bizari Coin de Carvalho^IV; Gilmar Fernandes do Prado^IV

Federal University of São Paulo (UNIFESP) São Paulo SP Brazil

^IDepartment of Neurology

^IIBrazilian Cochrane Center

^IIIDepartment of Emergency Medicine

^IVDepartment of Emergency Medicine, Department of Neurology

ABSTRACT

Convulsions triggered by fever are the most common type of seizures in childhood, and 20% to 30% of them have recurrence. The prophylactic treatment is still controversial, so we performed a systematic review to find out the effectiveness of continuous phenobarbital and intermittent diazepam compared to placebo for febrile seizure recurrence.

METHOD: Only randomized, double-blind, placebo-controlled trials were analyzed. The recurrence of febrile seizure was assessed for each drug.

RESULTS: Ten eligible clinical trials were included. Febrile seizure recurrence was smaller in children treated with diazepam or phenobarbital than in placebo group. Prophylaxis with either phenobarbital or diazepam reduces recurrences of febrile seizures. The studies were clinical, methodological, and statistically heterogeneous.

CONCLUSION: The effectiveness of phenobarbital and diazepam could not be demonstrated because clinical trials were heterogeneous, and the recommendation for treatment recurrence should rely upon the experience of the assistant physician yet.

Key words: prophylactic treatment, febrile seizures, intermittent diazepam, continuous phenobarbital.

RESUMO

As convulsões desencadeadas por febre são muito comuns na infância e 20% a 30% delas apresentam recorrência. O tratamento profilático, no entanto, ainda é controverso, motivo porque realizamos uma revisão sistemática para avaliar a eficácia do tratamento da recidiva de convulsão febril com diazepam e fenobarbital comparados a placebo.

MÉTODO: Analisamos somente estudos randomizados, duplo-cegos, controlados, utilizando fenobarbital contínuo ou diazepam intermitente versus placebo.

RESULTADOS: Dez ensaios clínicos foram incluídos. A recorrência de convulsão febril foi menor no grupo das crianças tratadas com diazepam ou fenobarbital em relação ao controle. Tanto diazepam quanto fenobarbital reduziram as recorrências da convulsão febril. Os estudos foram clínica, metodológica e estatisticamente heterogêneos.

CONCLUSÃO: A eficácia do fenobarbital e diazepam não pôde ser demonstrada nesta metanálise por causa da heterogeneidade dos ensaios clínicos, e a recomendação para tratamento de recorrência deve basear-se na experiência clínica de cada médico.

Palavras-chave: tratamento profilático, convulsões febris, diazepam intermitente, fenobarbital contínuo.

Febrile seizures are the most prevalent type of seizure, occurring in about 2% to 5% of young children, and their recurrence varies from 20% to 30%¹. Febrile seizure is defined as convulsions occurring in children aged 0 to 5 years, with no history of neuropsychiatric disorders or with a concomitant neurological disease. The prognosis is good for the most patients, but seizures are upsetting to the children and parents. There are studies demonstrating a relation between the number of febrile seizures, mainly the complex ones, and the risk for epilepsy^2-3. Besides, recurrent convulsions may be deleterious to intellectual development⁴.

The prevention of febrile seizures might be beneficial, but the long-term management is controversial^5-7. For some authors, prophylactic use of continuous phenobarbital prevents recurrence⁸. Others consider it of limited or no prophylactic value⁹. But they agree that side-effects and no compliance are common^3,9-10. Also intermittent diazepam prophylaxis has been used, but this treatment is not worldwide accepted and it has been argued that intermittent prophylaxis often fails¹¹.

A review of trials published in English showed that only phenobarbital or valproate has effect in preventing febrile seizures recurrence and besides the results, those drugs were not recommended because of adverse effects. We performed a systematic review including trials published in English and also in Spanish and Portuguese to analyze the recurrence of febrile seizure in children treated with phenobarbital/diazepam and placebo.

METHOD

The electronic search was performed combining the following key words: febrile convulsion, febrile seizures, convulsão febril, crisis febriles, convulsión febril, phenobarbital, fenobarbital, diazepam and prophylaxis, profilaxia, profilaxis. The reference research was obtained from Cochrane Center of Brazil, Medline, Lilacs and Embase. Other research sources were from letters to experts, thesis indexed at BIREME/PAHO-WHO (Biblioteca Regional Medicina/Panamerican Health Organization of the World Health Organization), abstracts sent to Medical Meetings, but not published. We also have searched over the reference list of all recovered trials.

Classification of the trials. We classified the trials according to the randomization in A (described as randomized, description presented), B (described as randomized, description absent) and C (described as randomized, description present contrary evidence). Taking in account the blinding procedure we classified the studies in A (described as double-blind, description present), B (described as double-blind, description absent) and C (described as double-blind, description present contrary evidence).

We included in this systematic review only placebo controlled trials, with all sample sizes, classified as A or B for randomization, A or B for blinding, written in English, Portuguese and Spanish. We excluded in this review case reports and trials classified as C for randomization and C for blinding. We analysed the treatment in three ways: intermittent diazepam compared to placebo; continuous phenobarbital compared to placebo; intermittent diazepam or continuous phenobarbital compared to placebo.

Statistical Analysis. The analysis was performed in table 2X2: febrile seizure recurrence (yes or no) and the intermittent use of diazepam (yes or no); febrile convulsion recurrence (yes or no) and the use of continuous phenobarbital (yes or no). The statistical heterogeneity of the results was evaluated in funnel plot graphics, related to sample size (Y axis) and calculation of c² test, in the N degrees of freedom, considering N = number of trials minus 1. To decide for a statistical significant finding it was used a<5%.

RESULTS

We identified forty-eight trials published in English related to continuous use of phenobarbital and intermittent use of diazepam in the treatment of recurrences of febrile seizures. Thirty-eigth trials were excluded, for they were not randomized ones. Ten randomized clinical trials in English were initially included in this review, for they fulfilled the inclusion criteria. Thirty-five trials in Portuguese and Spanish were found and among them only one was eligible.

Among those eleven trials^2,4,6,12-19 classified (Table 1), one trial (Knudsen¹²) was subsequently excluded from our statistical analysis, because of the classification C for randomization.

Diazepam versus placebo. Comparing the treatment with diazepam versus placebo (Figure 1), the recurrence risk in febrile convulsion is lower in children treated with intermittent diazepam (NNT 17; 95% CI 10 to 85; z=2.47 p<0.01). So for each 17 treated children, one has no recurrence (OR for recurrence 0.6; 95% CI 0.40 to 0.90). In the diazepam group 11.2% (44/393) have recurrence and also 17.1% (68/398) of the placebo group. The results comparing the trials with diazepam and placebo were heterogeneous (c² = 10.19; p<0.01).

Phenobarbital versus placebo. Comparing phenobarbital versus placebo (Fig 1), the recurrence risk is lower in children treated with continuous phenobarbital (NNT 8; CI 95% 5 to 18; z=3.46 p<0.05). If we treat 8 children with phenobarbital, comparing to placebo group, one child would not have recurrence (OR for recurrence 0.54; 95% CI 0.38 to 0.76). In the phenobarbital group 24.5% (71/290) have recurrence and also 37.0% (114/308) of the placebo group. The results comparing the trials with phenobarbital and placebo were heterogeneous (c²= 15.40 p<0.01).

Intermittent diazepam or phenobarbital versus placebo. Comparing the prophylatic effect of phenobarbital or diazepam with placebo (Fig 1), we found that phenobarbital or diazepam reduces recurrence risk in febrile convulsion (NNT 12; 95 % CI 8 to 22; c² = 25.76 p<0.02), meaning that for each 12 treated children, one has no recurrence. OR for recurrence is 0.56 (95% CI 0.43 to 0.73; z=-4.23 p<0.02).

DISCUSSION

The most common treatment for febrile seizures has been the daily administration of phenobarbital or intermittent diazepam^8,12, a practice that has been questioned by many authors. Since seventies, several studies concerning treatment of febrile seizures have been published, but only 10 trials could be considered in our analysis, disclosing a lack of good studies insuring a body of evidence to treat febrile seizures. Four studies (Table 2) comparing intermittent use of diazepam with placebo were found^2,16,18,19. Autret¹⁸, Mosquera¹⁶ and Rosman² demonstrated the efficacy of diazepam in preventing febrile seizure in relation to placebo, but the results in Autret and Mosquera studies were not statistically significant. Rosman demonstrated that the diazepam treatment was more effective than placebo in his study, but more than 20% were lost in follow-up, which could modify the result significantly. Uhari's trial¹⁹ showed no differences between the diazepam and placebo groups. The analysis of those four trials indicated that the risk for recurrent febrile seizures decreased with diazepam comparing to placebo, but the c² test demonstrates that those studies are heterogeneous, which means that the results concerning to efficacy of randomized trials with intermittent diazepam therapy are controversial.

Among the studies (Table 3) comparing use of phenobarbital and placebo, in five ones^4,6,13-15 phenobarbital reduced the recurrence risk for febrile seizures. But in 3 studies^4,13-15, the relative reduced risk was not significant. Mckinlay¹⁷ showed there was no difference in recurrence risk between phenobarbital and placebo.

The analysis of all studies (Table 4) in this meta-analytic review demonstrated that the treated group, either with intermittent diazepam or continuous phenobarbital, when compared with placebo group had a decreased recurrence risk of febrile seizure. The results were statistically significant, suggesting that the therapy is effective in preventing the febrile seizure recurrence. When we compared statistically the results, the c² test showed heterogeneity of the trials, either due to methodological differences or clinical intrinsic differences of each study, and three possibilities arise to explain this heterogeneity: answers that do not agree with the trial quality, sample size and even clinical heterogeneity. However, in view of methodological differences of each study design, it is not possible to make a decision concerning the efficacy of diazepam or phenobarbital in preventing febrile seizures based on this systematic review and mostly on the meta-analytic study.

Clinical heterogeneity detected in those included trials was mostly related to clinical definition of febrile seizure patients allocated to treatment, age of patients (below or above 5 years), number of seizures, duration of seizures, co-morbidities, abnormal EEG, and doses of medication in the same drug option. We did not find a study that have calculated the number of patients necessary to allow statistical power for a scientific decision on the effectiveness of the treatment.

In conclusion the present systematic review and meta-analyses allow us to conclude that there is not a strong recommendation to treat febrile seizure either with continuous phenobarbital or with intermittent diazepam prophylaxis, because of the design and heterogeneity of the primary studies, but there is an evidence of potential benefit with the treatment of both drugs, although it is not possible to conclude which therapy is better, due to already mentioned heterogeneity of study design. The treatment decision should result from appropriate judgement and experience of the physician, and a standard trial with a large number of patients should be done to bring more information over this issues.

Received 10 April 2003. Accepted 15 July 2003

Dr. Gilmar F. Prado - Rua Claudio Rossi 394 - 01547-000 São Paulo SP - Brasil. E-mail: [e-mail]

• 1. Nelson KB, Ellenberg JH . Prognosis in children with febrile seizures. Pediatrics 1978;61:720-727.
• 2. Rosman NP, Colton T, Labazzo J, et al. A controlled trial of diazepam administered during febrile illnesses to prevent recurrence of febrile seizures. N Engl J Med 1993;329:79-84.
• 3. Autret E, Ployet JL. Traitement des convulsions fébriles. Arch Pédiatr 2002;9:91-95.
• 4. Thilothammal N, Krishnamurthy PV, Kamala KG, Ahamed S, Banu K. Role of phenobarbitone in preventing recurrence of febrile convulsions. Indian Pediatrics 1993;30:637-642.
• 5. Millichap JG, Colliver JA. Management of febrile seizures: survey of current practice and phenobarbital usage. Pediatr Neurol 1991;7:243-248.
• 6. Farwell JR, Lee YJ, Hirtz DG, Sulzbacher SI, Ellenberg JH, Nelson KB. Phenobarbital for febril seizures: effects on intelligence and on seizure recurrence. N Engl J Med 1990;322:364-369.
• 7. Sulzbacher S, Farwell JR, Temkin N, Lu AS, Hirtz DG. Late cognitive effects of early treatment with phenobarbital. Clin Pediatr 1999;38:387-394.
• 8. Faero O, Kastrup KW, Lykkgaard-Nielsen E, Melchior JC, Thorn I. Successful prophylaxis of febrile convulsions with phenobarbital. Epilepsia 1972;13:279-285.
• 9. Heckmatt JZ, Houston AB, Clow DJ, et al. Failure of phenobarbitone to prevent febrile convulsions. BMJ 1976;76:559-561.
• 10. Newton RW. Randomized controlled trials of phenobarbitone and valproate in febrile convulsion. Arch Dis Child 1988;63:1189-1191.
• 11. Knudsen FU. Effective short-term diazepam prophylaxis in febrile convulsions. J Pediatr 1985;106:487-490.
• 12. Knudsen FU, Vestermark S. Prophylactic diazepam or phenobarbitone in febrile convulsions: a prospective controlled study. Arch Dis Child 1978;53:660-663.
• 13. Camfield PR, Camfield CS, Shapiro SH, Cummings C. The febrile seizure: antipyretic instruction plus either phenobarbital or placebo to prevent recurrence. J Pediatr 1980;97:16-21.
• 14. Bacon CJ, Muckow JC, Rawlins MD, Hierons AM. Placebo-controlled study of phenobarbitone and phenytoin in the prophylaxis of febrile convulsions. Lancet 1981;2:600-604.
• 15. Mamelle N, Mamelle JC, Plasse JC, Revol M, Gilly R. Prevention of recurrent febrile convulsions- a randomized therapeutic assay: sodium valproate, phenobarbital and placebo. Neuropediatr 1984;15:37-42.
• 16. Mosquera C, Rodriguez J, Cabrero A, et al. Prevención de la recurrencia de crisis febriles: profilaxis intermitente con diacepan rectal comparada con tratamiento continuo con valproato sódico. An Esp Pediatr 1987;27:379-381.
• 17. Mckinlay I, Newton R. Intention to treat febrile convulsions with rectal diazepam, valproate or phenobarbitone. Dev Med Child Neurol 1989;31:617-625.
• 18. Autret E, Billard C, Bertrand P, Motte J, Pouplard F, Joinville AP. Double-blind, randomized trial of diazepam versus placebo for prevention of recurrence of febrile seizures. J Pediatrics 1990;117:490-494.
• 19. Uhari M, Rantala H, Vainionpaa L, Kurtilla R. Effects of acetaminophen and of low intermittent doses of diazepam on prevention of recurrences of febrile seizures. J Pediatr 1995;126:991-995.

Publication Dates

• Publication in this collection
  06 Jan 2004
• Date of issue
  Dec 2003

History

• Accepted
  15 July 2003
• Received
  10 Apr 2003