INTERM ITTENT DIAZEPAM AND
CONTINUOUS PHENOBARBITAL TO TREAT
RECURRENCE OF FEBRILE SEIZURES
A system atic review w ith m eta-an alysis
Alice Hatsue Masuko
1, Aldemar Araujo Castro
2, Gustavo Rocha Santos
3,
Álvaro Nagib Atallah
3, Lucila Bizari Fernandes do Prado
4,
Luciane Bizari Coin de Carvalho
4, Gilmar Fernandes do Prado
4ABSTRACT - Convulsions trig g ered b y fever are the m ost com m on typ e of seizures in child hood , and 20% to 30% of them have recurrence. The p rop hylactic treatm ent is still controversial, so w e p erform ed a system atic review to find out the effectiveness of continuous p henob arb ital and interm ittent d iazep am com p ared to p laceb o for feb rile seizure recurrence. Method: Only rand om ized , d oub le-b lind , p laceb o-controlled trials were analyzed . The recurrence of feb rile seizure w as assessed for each d rug . Results: Ten elig ib le clinical trials were includ ed . Feb rile seizure recurrence was sm aller in child ren treated with d iazep am or p henob arb ital than in p laceb o g roup . Prop hylaxis with either p henob arb ital or d iazep am red uces recurrences of feb rile seizures. The st u d ies w ere clin ica l, m et h o d o lo g ica l, a n d st a t ist ica lly h et ero g en eo u s. Conclusion: Th e effect iven ess o f p henob arb ital and d iazep am could not b e d em onstrated b ecause clinical trials were heterog eneous, and the recom m end ation for treatm ent recurrence should rely up on the exp erience of the assistant p hysician yet.
KEY WORDS: p rop hylactic treatm ent, feb rile seizures, interm ittent d iazep am , continuous p henob arb ital.
Diazepam intermitente e fenobarbital contínuo para tratamento da recorrência de convulsões febris: uma revisão sistemática com metanálise
RESUMO - As convulsões desencadeadas por febre são m uito com uns na infância e 20% a 30% delas apresentam recorrência. O tratam ento p rofilático, no entanto, aind a é controverso, m otivo p orq ue realizam os um a revisão sistem ática p ara avaliar a eficácia d o tratam ento d a recid iva d e convulsão feb ril com d iazep am e fenob arb ital com p arad os a p laceb o. Método: An alisam o s so m en te estu d o s ran d o m izad o s, d u p lo -ceg o s, co n tro lad o s, utilizand o fenob arb ital contínuo ou d iazep am interm itente versus p laceb o. Resultados: Dez ensaios clínicos foram incluíd os. A recorrência d e convulsão feb ril foi m enor no g rup o d as crianças tratad as com d iazep am ou fen o b arb it al em relação ao co n t ro le. Tan t o d iazep am q u an t o fen o b arb it al red u ziram as reco rrên cias d a convulsão febril. Os estudos foram clínica, m etodológica e estatisticam ente heterogêneos. Conclusão: A eficácia d o fenob arb ital e d iazep am não p ôd e ser d em onstrad a nesta m etanálise p or causa d a heterog eneid ad e d os ensaios clínicos, e a recom end ação p ara tratam ento d e recorrência d eve b asear-se na exp eriência clínica d e cad a m éd ico.
PALAVRAS-CHAVE: tratam ento p rofilático, convulsões feb ris, d iazep am interm itente, fenob arb ital contínuo.
Feb rile seizu res are t h e m o st p revalen t t yp e o f seizu re, o ccu rrin g in ab o u t 2% to 5% o f yo u n g ch il-d ren , an il-d t h eir recu rren ce varies fro m 20% t o 30%1. Feb rile seizu re is d efin ed as co n vu lsio n s o ccu rrin g in ch ild ren ag ed 0 to 5 years, w ith n o h isto ry o f n eu ro p sych iatric d iso rd ers o r w ith a co n co m itan t n eu ro -lo g ical d isease. Th e p ro g n o sis is g o o d fo r th e m o st
p atien ts, b u t seizu res are u p settin g to th e ch ild ren an d p aren t s. Th ere are st u d ies d em o n st rat in g a re-lation between the num ber of febrile seizures, m ainly the com plex ones, and the risk for epilepsy2-3. Besides, recu rren t co n vu lsio n s m ay b e d eleterio u s to in tel-lect u al d evelo p m en t4.
The prevention of febrile seizures m ight be
benefi-Fed eral Un iversit y o f São Pau lo (UNIFESP) São Pau lo SP Brazil: 1Dep art m en t o f Neu ro lo g y; 2Brazilian Co ch ran e Cen t er; 3Dep art m en t o f
Em erg en cy Med icin e an d 4Dep art m en t o f Em erg en cy Med icin e, Dep art m en t o f Neu ro lo g y.
Received 10 Ap ril 2003. Accep t ed 15 Ju ly 2003.
t h ey ag ree t h at sid e-effect s an d n o co m p lian ce are com m on3,9-10. Also interm ittent diazepam prophylaxis h as b een u sed , b u t th is treatm en t is n o t w o rld w id e accep t ed an d it h as b een arg u ed t h at in t erm it t en t p ro p h ylaxis o ft en fails11.
A review o f trials p u b lish ed in En g lish sh o w ed t h at o n ly p h en o b arb it al o r valp ro at e h as effect in p reven t in g feb rile seizu res recu rren ce an d b esid es t h e resu lt s, t h o se d ru g s w ere n o t reco m m en d ed b ecau se o f ad verse effect s. We p erfo rm ed a syst e-m atic review includ ing trials p ub lished in English and also in Sp an ish an d Po rt u g u ese t o an alyze t h e recu r-ren ce o f feb rile seizu re in ch ild r-ren t reat ed w it h p h e-n o b arb it al/d iazep am ae-n d p laceb o .
M ETHOD
Th e elect ro n ic search w as p erfo rm ed co m b in in g t h e fo llo w in g key w o rd s: feb rile co n vu lsio n , feb rile seizu res, co n vu lsão feb ril, crisis feb riles, co n vu lsió n feb ril, p h en o -b ar-b it al, fen o -b ar-b it al, d iazep am an d p ro p h ylaxis, p ro fila-xia, p ro filaxis. Th e referen ce research w as o b t ain ed fro m Co ch ran e Cen t er o f Brazil, Med lin e, Lilacs an d Em b ase. Oth er research so u rces w ere fro m letters to exp erts, th esis in d exed at BIREME/PAHO-WHO (Bib lio t eca Reg io n al Med i-cina/Panam erican Health Organization of the World Health Organization), ab stracts sent to Med ical Meetings, b ut not p u b lish ed . We also h ave search ed o ver t h e referen ce list o f all reco vered t rials.
Classification of the trials. We classified the trials accor-d in g t o t h e ran accor-d o m izat io n in A (accor-d escrib eaccor-d as ran accor-d o m izeaccor-d , d escrip t io n p resen t ed ), B (d escrib ed as ran d o m ized , d crip t io n a b sen t ) a n d C (d escrib ed a s ra n d o m ized , d es-crip t io n p resen t co n t rary evid en ce). Takin g in acco u n t t h e b lind ing p roced ure we classified the stud ies in A (d escrib ed as d oub le-b lind , d escrip tion p resent), B (d escrib ed as d ou-b ou-b lin d , d escrip t io n aou-b sen t ) an d C (d escriou-b ed as d o u ou-b le-b lin d , d escrip t io n p resen t co n t rary evid en ce).
We includ ed in this system atic review only p laceb o con-t ro lled con-t rials, w icon-t h all sam p le sizes, classified as A o r B fo r rand om ization, A or B for b lind ing , w ritten in Eng lish, Po r-t u g u ese an d Sp an ish . We exclu d ed in r-t h is review case re-p o rt s an d t rials classified as C fo r ran d o m izat io n an d C fo r b lin d in g . We an alysed t h e t reat m en t in t h ree w ays: in t er-m it t en t d iazep aer-m co er-m p ared t o p laceb o ; co n t in u o u s p h e-n o b arb it al co m p ared t o p laceb o ; ie-n t erm it t ee-n t d iazep am o r co n t in u o u s p h en o b arb it al co m p ared t o p laceb o .
Statistical Analysis. The analysis was perform ed in table
2X2: feb rile seizu re recu rren ce (yes o r n o ) an d t h e in t er-m it t en t u se o f d iazep aer-m (yes o r n o ); feb rile co n vu lsio n recu rren ce (yes o r n o ) an d t h e u se o f co n t in u o u s p h en o
-w as u sed α< 5%.
RESULTS
We id e n t ifie d fo rt y-e ig h t t ria ls p u b lish e d in En g lish related to co n tin u o u s u se o f p h en o b arb ital an d in t erm it t en t u se o f d iazep am in t h e t reat m en t o f recu rren ces o f feb rile seizu res. Th irt y-eig t h t rials w ere exclu d ed , fo r th ey w ere n o t ran d o m ized o n es. Ten rand om ized clinical trials in Eng lish w ere initially includ ed in this review, for they fulfilled the inclusion crit eria. Th irt y-five t rials in Po rt u g u ese an d Sp an ish w ere fo u n d an d am o n g th em o n ly o n e w as elig ib le. Am o n g th o se eleven trials2,4,6,12-19 classified (Tab le 1), o n e t rial (Kn u d sen12) w as su b seq u en t ly exclu d ed fro m o u r st at ist ical an alysis, b ecau se o f t h e classifi-catio n C fo r ran d o m izatio n .
Diazepam versus placebo. Co m p arin g th e treat-m en t w it h d iazep atreat-m versus p laceb o (Fig u re 1), t h e recu rren ce risk in feb rile co n vu lsio n is lo w er in ch il-d ren t reat eil-d w it h in t erm it t en t il-d iazep am (NNT 17; 95% CI 10 t o 85; z= 2.47 p < 0.01). So fo r each 17 treated child ren, one has no recurrence (OR for recur-ren ce 0.6; 95% CI 0.40 t o 0.90). In t h e d iazep am g ro u p 1 1 .2 % (4 4 /3 9 3 ) h a ve recu rren ce a n d a lso 17.1% (68/398) o f t h e p laceb o g ro u p . Th e resu lt s com p aring the trials with d iazep am and p laceb o we-re h etero g en eo u s (χ2 = 10.19; p < 0.01).
Phenobarbital versus placebo. Co m p arin g p h e-nob arb ital versus p laceb o (Fig 1), the recurrence risk is lo w er in ch ild ren treated w ith co n tin u o u s p h en o -b ar-b it al (NNT 8; CI 95% 5 t o 18; z= 3.46 p < 0.05). If w e treat 8 ch ild ren w ith p h en o b arb ital, co m p arin g to p laceb o g ro u p , o n e ch ild w o u ld n o t h ave recu r-ren ce (OR fo r recu rr-ren ce 0.54; 95% CI 0.38 t o 0.76). In t h e p h en o b arb it al g ro u p 24.5% (71/290) h ave recu rren ce an d also 37.0% (114/308) o f t h e p laceb o g ro u p . Th e resu lts co m p arin g th e trials w ith p h e-n o b arb it al ae-n d p laceb o w ere h et ero g ee-n eo u s (χ2 = 15.40 p < 0.01).
DISCUSSION
Th e m o st co m m o n t reat m en t fo r feb rile seizu res h as b een t h e d aily ad m in ist rat io n o f p h en o b arb it al or interm ittent d iazep am8,12, a p ractice that has b een
q uestioned b y m any authors. Since seventies, several stud ies concerning treatm ent of feb rile seizures have b een p u b lish ed , b u t o n ly 10 t ria ls co u ld b e co n sid ered in o u r an alysis, d isclo sin g a lack o f g o o d stu
-Fig 1. Odds Rartio (OR) and 95%CI in relation to number of recurrences w ith treatment Drug vs Placebo.
Table 1. Characteristics of the trials included in this review.
Referen ce Year Ran d o m izat io n Do u b le- Sam p le Lo ss
Au t h o r Nu m b er b lin d (N) Exclu sio n s N %
Kn u d sen 12 1978 C - 195 0 39 20
Cam field 13 1980 B A 79 0 12 15,2
Baco n 14 1981 B A 207 0 69 33,3
Mam elle 15 1984 A - 49 1 3 6,1
Mo sq u era 16 1987 A - 43 0 4 9,3
Mckin lay 17 1989 B - 101 0 13 12,9
Au t ret 18 1990 B A 185 0 18 9,7
Farw ell 6 1990 A A 217 0 44 20.3
Ro sm an 2 1993 A A 406 0 105 25,8
Th ilo th am m al 4 1993 A A 60 0 3 5
d ies in su rin g a b o d y o f evid en ce t o t reat feb rile sei-zu res. Fo u r st u d ies (Tab le 2) co m p arin g in t erm it t en t u se o f d iazep am w it h p laceb o w ere fo u n d2,16,18,19. Autret18, Mosquera16 and Rosm an2 dem onstrated the efficacy o f d iazep am in p reven t in g feb rile seizu re in relat io n t o p laceb o , b u t t h e resu lt s in Au t ret an d Mo sq u era st u d ies w ere n o t st at ist ically sig n ifican t . Rosm an d em onstrated that the d iazep am treatm ent w as m o re effective th an p laceb o in h is stu d y, b u t m o re th an 20% w ere lo st in fo llo w -u p , w h ich co u ld m o d ify t h e resu lt sig n ifican t ly. Uh ari‘s t rial19 show ed n o d ifferen ces b et w een t h e d iazep am an d p laceb o g ro u p s. Th e an alysis o f t h o se fo u r t rials in d icat ed t h at t h e risk fo r recu rren t feb rile seizu res d ecreased with d iazep am com p aring to p laceb o, b ut the χ2 test d em onstrates that those stud ies are heterog eneous, w h ich m ean s t h at t h e resu lt s co n cern in g t o efficacy of rand om ized trials with interm ittent d iazep am the-rap y are co n t ro versial.
Am o n g t h e st u d ies (Tab le 3) co m p arin g u se o f p h en o b arb it al an d p laceb o , in five o n es4,6,13-15 p he-nob arb ital red uced the recurrence risk for feb rile sei-zu res. Bu t in 3 st u d ies4,13-15, t h e relat ive red u ced risk
w as n o t sig n ifican t . Mckin lay17 sh o w ed t h ere w as n o d ifferen ce in recu rren ce risk b et w een p h en o b ar-b it al an d p lacear-b o .
Th e an alysis o f all st u d ies (Tab le 4) in t h is m et a-analytic review dem onstrated that the treated group, either with interm ittent diazepam or continuous phe-nob arb ital, when com p ared with p laceb o g roup had a d ecreased recu rren ce risk o f feb rile seizu re. Th e resu lt s w ere st at ist ically sig n ifican t , su g g est in g t h at t h e t h erap y is effect ive in p reven t in g t h e feb rile sei-zure recurrence. When w e com p ared statistically the results, the χ2 test showed heterogeneity of the trials, eit h er d u e t o m et h o d o lo g ical d ifferen ces o r clin ical in t rin sic d ifferen ces o f each st u d y, an d t h ree p o ssi-b ilit ies arise t o exp lain t h is h et ero g en eit y: an sw ers t h at d o n o t ag ree w it h t h e t rial q u alit y, sam p le size an d even clin ical h etero g en eity. Ho w ever, in view o f m et h o d o lo g ical d ifferen ces o f each st u d y d esig n , it is n o t p o ssib le to m ake a d ecisio n co n cern in g th e efficacy o f d iazep am o r p h en o b arb it al in p reven t in g feb rile seizu res b ased o n t h is syst em at ic review an d m o st ly o n t h e m et a-an alyt ic st u d y.
Clin ical h etero g en eity d etected in th o se in clu d ed
Ro sm an 0.33 8/8 23 OA 23
Mo sq u era 0.5 8/8 24 RA 24
Au t ret 0.2 –0.5 12/12 10.3 OA 10.3
Uh ari 0.2 –0.65 8/8 24 RA 24
RA, rect al ad m in ist rat io n ; AO, o ral ad m in ist rat io n d u rin g t h e feb rile ep iso d e
Table 3. Trials using phenobarbital in relation to dose, duration of treatment, administration and follow -up.
Ph en o b arb it al
Au t h o r Do se Du rat io n o f Ad m in ist rat io n Fo llo w -u p
(m g /Kg /d ay) t reat m en t (m o n t h s) (m o n t h s)
Th ilo th am m al 5 12 AO 12
McKin lay 5 3 - 6 AO 24
Farw ell 4 – 5 24 AO 24
Mam elle 3 – 4 21 AO 36
Cam field 4 – 5 12 AO 12
Bacon 5 12 AO 12
t ria ls w a s m o st ly rela t ed t o clin ica l d efin it io n o f feb rile seizu re p at ien t s allo cat ed t o t reat m en t , ag e o f p a t ien t s (b elo w o r a b o ve 5 yea rs), n u m b er o f se izu re s, d u ra t io n o f se izu re s, co -m o rb id it ie s, ab n o rm al EEG, an d d o ses o f m ed icat io n in t h e sam e d ru g o p t io n . We d id n o t fin d a st u d y t h a t h a ve calculated the num b er of p atients necessary to allow st a t ist ica l p o w e r fo r a scie n t ific d e cisio n o n t h e effect iven ess o f t h e t reat m en t .
In co n clu sio n th e p resen t system atic review an d m et a-an alyses allo w u s t o co n clu d e t h at t h ere is n o t a st ro n g reco m m en d a t io n t o t rea t feb rile seizu re e it h e r w it h co n t in u o u s p h e n o b a rb it a l o r w it h in t erm it t en t d iazep am p ro p h ylaxis, b ecau se o f t h e d esig n and heterog eneity of the p rim ary stud ies, b ut t h ere is an evid en ce o f p o t en t ial b en efit w it h t h e treatm en t o f b o th d ru g s, alth o u g h it is n o t p o ssib le t o co n clu d e w h ich t h erap y is b et t er, d u e t o alread y m e n t io n e d h e t e ro g e n e it y o f st u d y d e sig n . Th e t reat m en t d ecisio n sh o u ld resu lt fro m ap p ro p riat e ju d g em en t an d exp erien ce o f t h e p h ysician , an d a stand ard trial with a large num b er of p atients should b e d o n e to b rin g m o re in fo rm atio n o ver th is issu es.
REFERENCES
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 ef-fects 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. Epi-lepsia 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. New ton 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 compara-da 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, valpro ate o r pheno barbito ne. Dev Med Child Neuro l 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.
Table 4. Febrile seizure recurrence regarding to phenobarbital or diazepam compared to placebo.
Au t h o r Referen ce Year N = ch ild ren w it h Ph en o b arb it al o r Diazep am
Nu m b er Ph en o b arb it al Diazep am Placeb o
n /N n /N n /N
Cam field 13 1980 2 / 39 10 / 40
Baco n 14 1981 10 / 48 15 / 43
Mam elle 15 1984 4 / 24 9 / 26
Mo sq u era 16 1987 1 / 18 4 / 25
Mckin lay 17 1989 12 / 41 14 / 60
Farw ell 6 1990 41 / 108 50 / 109
Au t ret 18 1990 15 / 93 18 / 92
Ro sm an 2 1993 7/ 202 29 / 204
Th ilo t h am m al 4 1993 2 / 30 16 / 30
Uh ari 19 1995 21 / 80 17 / 77
