ArchivesofDiseaseinChildhood 199470:395-399 [605252]

ArchivesofDiseaseinChildhood 1994;70:395-399
Familyhistoryandrecurrence offebrileseizures
AvanEsch,EWSteyerberg, MYBerger,MOffringa, GDerksen-Lubsen,JDFHabbema
Abstract
Todetermine thevalueofadetailed
familyhistory fortheassessment ofthe
riskofrecurrence offebrileseizures, 115
children whovisitedtheemergency room
ofanacademic children's hospital were
studied prospectively. Therecurrence
riskoffebrile seizures wasanalysed in
relation tothechild's family history
andtheproportion ofrelatives affected
byfebrile seizures usingKaplan-Meier
estimates andCoxproportional hazard
models. Afirstdegree family history
positive forfebrile seizures (parents or
siblings affected byfebrile seizures)
increased achild's twoyearrecurrence
riskfrom27to52%.Nosignificant
increase ofrecurrence riskforfebrile
seizures wasfound inchildren with
second degree relatives (grandparents
anduncles/aunts) orcousins only
affected byfebrile seizures. Recurrence
riskwassignificantly correlated with
theproportion offirstdegree relatives
affected byfebrileseizures: riskswere27,
40,and83%inchildren whoseproportion
was0,0-0-5,andn05respectively.
Analysis oftherecurrence riskinrelation
toaweighted proportion, adjusted forthe
attained ageandsexoffirstdegree
relatives, showed similar results. It is
concluded thattheapplication ofthe
proportion offirstdegree relatives
affected byfebrile seizures generates a
moredifferentiated assessment ofthe
recurrence riskoffebrileseizures.
(ArchDisChild1994;70:395-399)
Thecumulative incidence offebrileseizures
inchildren inEuropeisbetween 2and50/o.l-3
Onaverage, 30%ofchildren haveasecond
febrileseizureand15%havetwoormorerecur-
rencesaftertheirinitialfebrileseizure.3-7
Although manyworkers advocate prophylaxis
fortherecurrence offebrileseizures, con-
troversyexistsaboutthetreatment ofchoice.8'2
Moreover, itisstillnotpossible todiscriminate
between children whowillandchildren who
willnothavearecurrence. Furtherimprove-
mentsintheabilitytopredicttherecurrence of
afebrileseizurewillaiddoctorsinchoosing the
appropriate prophylactic treatment, ifany.
Afirstdegree-familyhistorypositive for
febrileseizures hasbeenshowntobeamajor
riskfactorfortherecurrence ofafebrileseizure
inseveralstudies; otherriskfactorsareyoung
ageatonset,multiple initialseizures, andrela-
tivelylowtemperature attheinitialseizure.47 13
Therecurrence risksforfebrileseizures have
beenreported inrelation tothepresence offirstdegreerelatives (parents andsiblings)
affected byfebrileseizures orinrelation
tothepresence ofaffected relatives ofany
degree.3 1415Thepredictive valueofthe
presence ofaffected secondandthirddegree
relatives ontherecurrence riskoffebrile
seizures isunknown. Also,thenumber ofa
child'srelatives hasnotsofarbeentakeninto
account. Ingeneral, achild'schanceofhaving
afamilyhistorypositive forfebrileseizures
willbeproportional tothenumber ofrelatives.
Thuschildren withlargerfamilies willbemore
likelytohaveapositive familyhistory.'6 An
incorporation ofthenumber ofrelatives in
thefamilyhistoryoffebrileseizures mayyield
amoreaccurate assessment ofachild's
recurrence risk.
Weinvestigated theassociation between the
recurrence offebrileseizures andthepresence
ofaffected firstdegree relatives andthe
presence ofaffected second (grandparents,
uncles/aunts) orthirddegree (cousins)
relatives separately. Wealsoinvestigated the
recurrence offebrileseizures inchildren
inrelation totheproportion offirstdegree
relatives affected.
Patients andmethods
Inanongoing prospective clinicbasedfollow
upstudy142consecutive children withan
initialfebrileseizureatbetween 6monthsand
6yearsofagewereincluded. Theyattended
theemergency roomoftheSophiaChildren's
Hospital/Academic Hospital ofRotterdam
between February 1988andFebruary 1990.
Febrile seizures weredefined inaccordance
withtheNational Institute ofHealth con-
sensusstatement.17 Feverhadtobevalidated
athomeorinthehospital asarectal
temperature of3850Corgreaterwithina
periodoftwohoursbeforeuntiltwohours
afterseizure occurrence. Arecurrence of
febrileseizure wasdefined asasubsequent
febrileseizureduring anewfebrileperiod.
Ageatonset,gender, parental country of
origin,seizuretype(duration, generalisation,
multiplicity), temperature atonset,andfirst
degreefamilyhistoryoffebrileseizures were
recorded onstandard formsatthefirstvisit.
Children withremaining neurological damage
orsubsequent afebrileseizures(threechildren)
andchildren givencontinuous prophylaxis
(phenobarbitone orsodium valproate) for
morethanthreemonths (14children) were
excluded, leavingastudygroupof125.
Parents wereaskedtoreporttherecurrences
offebrile seizures totheinvestigators.
Recurrence histories wereascertained atfollow
upvisitstotheclinicbyoneoftheauthors
(MO).Recurrence datesandcharacteristicsDepartment ofPediatrics, Academic
HospitalRotterdam/Sophia
Children's Hospital,
TheNetherlands
AvanEsch
MOffringaGDerksen-Lubsen
CenterforClinical
Decision Sciences,
Department ofPublicHealth, ErasmusUniversity,Rotterdam, The
Netherlands
AvanEsch
EWSteyerberg
MYBerger
MOffiinga
JDFHabbema
Correspondence to:
DrArjenvanEsch,Center
forClinicalDecisionSciences, Department ofPublicHealth, Room
Ee2091,Erasmus University,FacultyofMedicine, Post
OfficeBox1738,3000DR
Rotterdam, The
Netherlands.
Accepted 5January 1994395 on February 26, 2020 by guest. Protected by copyright. http://adc.bmj.com/ Arch Dis Child: first published as 10.1136/adc.70.5.395 on 1 May 1994. Downloaded from

vanEsch,Steyerberg, Berger,Offringa,Derksen-Lubsen, Habbema
wererecorded onstandard forms.Subsequent
toamailedannouncement twoyearsafterthe
initialfebrileseizure, oneinvestigator (AvE)
contacted theparents byphonetoobtain
complete recurrence ascertainment. The
parentsof10children couldnotbecontacted.
Oftheremaining 115children, adetailed
historyofallfirstdegreerelatives (parents and
siblings) -thatis,birthdate,sex,andseizure
history (febrile ornot,date,cause) -was
obtained fromtheparents. InWestEuropean
children adetailed historywasalsoobtained of
allseconddegreerelatives (grandparents and
uncles/aunts) andpartofthirddegreerelatives
(cousins). Innon-West European children a
detailed history wasobtained ofallfirst
degreerelatives, butonlyofthosesecondand
thirddegreerelatives whowereaffected by
febrileseizures. Detailed familyhistorydata
wererecorded onstandard forms.Seizures in
relatives whichoccurred aftertheinitial
seizureoftheindexchildwerenottakeninto
account. Wheretherewasuncertainty abouta
relative's history,parentswereaskedtocollect
additional information andwerecontacted
oncemoreatalaterdate.Relatives whose
febrileseizurehistoryremained unknown were
nottakenintoaccount intheanalysis.
Twoyearcumulative risksofoneandtwo
recurrences offebrileseizures wereestimated
withKaplan-Meier survival analysis.'8
Univariate andmultivariate Coxproportional
hazard regression models wereusedto
examine theeffectofriskfactors onthe
probability ofsubsequent febrileseizures. 19
Hazard ratioswith95%/oconfidence intervals
(CIs)werecomputed tocompare risksof
different patientsubgroups. Thehazardratio
maybeinterpreted asarelative riskof
recurrence.
Firstly,recurrence risksforfebrileseizures
wereanalysed inrelation tothepresence of
firstdegreerelatives affected byfebrileseizures
andinrelation tothepresence ofaffected
seconddegreerelatives orcousins. Differences
.0
-00~
a)
E
C)
Age(months)
Figure1 Cumulative probability ofhavingafirstfebrile
seizurebefore acertain agebased onOffringa etal.The
probability at72months issetatunity.Thevalues are
usedasweightsforrelatives inrelation totheirgenderand
age.inthepresence ofrelatives affected byfebrile
seizures between children wereanalysed with
Pearson's x2test.
Secondly, recurrence riskswereanalysed in
relation totheproportion offirstdegree
relatives affected byfebrileseizures, excluding
theindexchild.Thisproportion wascalledthe
crudeproportion ofrelatives affected byfebrile
seizures andcanbeexpressed as
Naffected/Ntotal
Forexample, thecrudeproportion wouldbe
033ifachildhadoneunaffected siblingand
twoparents, oneofwhomwasaffected.
Finally, riskswereanalysed inrelation
toaweighted proportion. Thisweighted
proportion adjustsforthelowerprobability of
apositive seizurehistory inyoungsiblings.
Weights forrelatives wereestimated fromthe
cumulative probability distribution ofageat
onsetinchildren withfebrileseizures ina
population basedstudy'(fig1)-forexample,
aboy20monthsofagewasassigned aweight
of059.Allparentsandchildren of6yearsof
ageorolderwereassigned aweightofunity.
Thedenominator oftheweighted proportion
constitutes thesummated weights (W)of
allfirstdegreerelatives (n),whereas the
numerator constitutes thenumber ofpositive
relatives (Naffected), asitisinthecrudepropor-
tion.Thustheweighted proportion ofrelatives
affected byfebrileseizures canbeexpressed as
Naffected/XWi (i-l..n)
Amultivariate Coxproportional hazards
modelwasusedtoexamine theeffectoffamily
history, thecrudeproportion, andthe
weighted proportion ofrelatives affected by
febrileseizures ontheprobability ofthe
recurrence offebrileseizures, adjusting for
otherpublished riskfactorsfortherecurrence
offebrileseizures.4-7 13Thesefactorswereage
atonset(divided inthreesubgroups: <1,
1-25,and>2-5years),seizuretype(simpleor
multiple), andtemperature atthetimeofthe
firstseizure(moreorlessthan40°C).The
effectofageonrecurrence riskwasanalysed
withthelogranktestfortrend.
Results
Onehundred andfifteen children were
included. SixtyfivewereofWestEuropean
origin(mainlyDutch) and50wereofnon-
WestEuropean origin(mainly Mediterranean
andCaribbean). Themeanageatthefirst
febrileseizure was1-7years.Themedian
followupofchildren without recurrences
was2-1years;81(700/o)werefollowed for
morethantwoyears.Thirtysix(31O%)children
hadonerecurrence and18hadtworecur-
rences; thetwoyearrisksforoneandtwo
recurrences were31and16%respectively (fig
2).Table1givestheclinicalcharacteristics of
the115children withthenumber, percentage,
andhazardratiosofchildren withrecurrences.
Recurrence risksweresignificantly increased in
children withmultiple initialseizures(hazard396 on February 26, 2020 by guest. Protected by copyright. http://adc.bmj.com/ Arch Dis Child: first published as 10.1136/adc.70.5.395 on 1 May 1994. Downloaded from

Familyhistoryandrecurrence offebrieseizures
children withaffected firstdegreerelatives,
however, weresignificantly increased (uni-
variatehazardratio2-5;multivariate hazard
ratio3-2).Risksoftworecurrences showed a
similarresult(table2).
Recurrence riskswerestudied inrelation to
thepresence ofaffected parents orthepres-
enceofaffected siblings separately. Table3
showsthattherecurrence riskwasincreased
l 1Recurrence from28to62%whenthechildhadanaffected
parent(hazardratio3-1;CI14to67).When
~2Recurrences asiblingaffected byfebrileseizures was
_>2Recurrences presenttherecurrence riskwasincreased from
29to55°/O(hazard ratio2A4;CI10to5-8).
Risksoftworecurrences wereincreased from
0~0151101'52102 514to31%(hazardratio2X7;CI09to81)and
from15to27%(hazardratio2-2;CI0-6toYearssincefirstseizure 7-7)respectively. Thusthepresence ofparents
Probability ofoneandtworecurrences afterfirstaffected byfebrileseizures andofsiblingsizure affected byfebrileseizures hadsimilareffects
ontherecurrence risk.
-3),initialseizuredurations ofmore Recurrence risksofWestEuropean and
5minutes (hazardratio2-3),andarela-non-West European children werestudied
lowtemperature attheinitialseizure separately because thepercentage ofchildren
Iratio2-1).Ageatonset(p=0-07, trend withfirstdegreerelatives affected byfebrile
render,andoriginofparentshadnoseizures innon-West European children
anteffectonrecurrence risk. (10%)waslowerthaninWestEuropean
children (25%;p=004).InWestEuropean
children recurrence riskswere19%inthose
HISTORY OFFEBRILE SEIZURES without anyaffected firstdegreerelatives
iledfebrileseizurehistoryof227(totaland50%inthosewithaffected firstdegree
Lrentsand121(total122)siblingscould relatives. Innon-West European children,
ained.Thirteen (6%)parents and12recurrence riskswere35and60%respec-
siblings had had afebrileseizure. tively.Univariate hazardratiosinchildren
-en(2-8%) of610recorded second withaffected firstdegreerelatives weresimilar
relatives and12(2-4%)of493recorded -thatis,2-8(CI11to71)inWestEuropean
legreerelatives ofchildren ofWestchildren and33(CI09to116)innon-West
anoriginhadhadafebrileseizure. InEuropean children.
total,24seconddegreerelatives and20
cousinshadhadafebrileseizure.
Twenty one(18%)children hadaffected
firstdegreerelatives (table2).Twenty five
(22%)children hadaffectedgrandparents or
uncles/aunts (second degreerelatives) or
cousins(partofthirddegreerelatives). Risks
ofoneandtworecurrences inchildren with
affected seconddegreerelatives orcousins
weresimilar totherisksofchildren without
anyaffectedrelative. Risksofonerecurrence in
Table1Recurrence risksinrelationtoclinicalcharacteristics
Noatrisk Nowithseizure Hazardratio
Feature (n=115) recurrence (risk*) (CI)t
Ageatonset(years)
<1 27 11(41) 1-4(07to28)
1-25 72 23(32) rc
>25 16 2(13) 04(0-1to1-5)Gender
Male 71 23(35) 11(0-5to2-1)
Female 44 13(27) rcOriginofparents
WestEuropean 65 18(27) rc
Non-West European 50 18(37) 1-3(0-7to2-5)
SeizuretypeMultiple 29 14(49) 2-3(1-2to45)Simple 86 22(26) rcGeneralised 108 34(32) rc
Focal 7 2(29) 0-9(0-2to3-7)l15minutes 16 8(53) 2-3(1-0to50)
<15minutes 99 28(28) rcTemperature atonset-400C 59 13(23) rc<400C 56 23(41) 2-1(I*1to4-1)
*Kaplan-Meier estimates(%/6)oftwoyearcumulative incidence.tUnivariate hazardratioswith95%CIcompared withreference catagory (rc).CRUDE PROPORTION OFRELATIVES AFFECTED
BYFEBRILE SEIZURES
Analysis ofrecurrences inrelation tothecrude
proportion ofaffected relatives showed thatthe
riskofonerecurrence increased whenthe
proportion increased (seefig3andtable4).The
riskoftwoormore recurrences wasonly
increased inchildren withaproportion greater
than orequal to05(table4),however.
Univariate hazardratiosinthisgroup were6-3
foroneand5-2fortworecurrences respectively;
multivariate hazardratios were6-8and5-7.
WEIGHTED PROPORTION OFRELATIVES
AFFECTED BYFEBRILE SEIZURES
Finally, recurrence risks wereanalysed in
relation totheweighted proportion ofaffected
relatives, whichisadjusted fortheattained age
andsexoftherelatives (table4).Although
risksandhazardratios weregenerally lower,
thesamepattern wasobserved asdescribed in
thecrudeproportion ofrelatives affected by
febrileseizures.
Discussion
Theaimofthisstudy wastodetermine the
valueofadetailed familyhistoryforthepre-
dictionoftherecurrence offebrileseizures. In100
90
80
70
60
50
40
30
20
10
0a)
a)C.)
CTa)
.)
0
Q
Figure2
febrilesei
ratio2
than15
tively1
(hazard
test),gsignific~
FAMILY
Adetai
230)pa
beobtu
(10%)
Seventi
degree.
thirdd
Europe397 on February 26, 2020 by guest. Protected by copyright. http://adc.bmj.com/ Arch Dis Child: first published as 10.1136/adc.70.5.395 on 1 May 1994. Downloaded from

vanEsch,Steyerberg, Berger,Offringa, Derksen-Lubsen, Habbema
Table2Numberofrecurrences offebrileseizures, risksandhazardratiosinrelationtofamilyhistoryoffebrile seizure
1Recurrence a2Recurrences
Hazardratio(CI)t Hazardratio(CI)t
Familyhistory NoatNo Nooffebrileseizures risk (risk*) Univariate Multivariate (risk*) Univariate Multivariate
None 69 19(27) 10(15)
Seconddegree/cousins 25 6(24) rc rc 3(12) rc rc
Firstdegreerelatives 21 11(52)2-5(1-2to51)3-2(1-6to66)5(23)2-0(07to55)2-2(07to5-8)
All 115 36(31) 18(16)
*Kaplan-Meier estimates(%/6)oftwoyearcumulative incidence.tHazardratioswith95%CIcompared withreference category(rc).
thisprospective followupstudytherecurrence
riskswerestudiedinrelationtothepresence of
relatives affected byfebrileseizures among
first,second, andapartofthirddegree
relatives, andinrelation totheproportion of
relatives affectedbyfebrileseizures.
Detailed familyhistorydatawereobtained
throughinterviews byphone.Thismethodhas
beenshowntobealmostasaccurate asdirect
interviews.20 Therefore weassume the
accuracy ofourinterviews tobeequivalent
tothatofapaediatrician's orgeneralpracti-
tioner's interview. History dataforfebrile
seizuresofseconddegreerelatives andcousins
wereobtained uptotwoyearsaftertheinitial
seizureinsomechildren, possiblyintroducing
somerecallbias.Inourexperience, however,
verylittleadditional information onfamily
historybecomes available afterseizurerecur-
rences.Mostfamilyinformation isgathered by
theparentsaftertheoccurrence oftheirchild's
initialseizure.Thustheaccuracy ofdataonthe
historyoffebrileseizures infirstdegreerela-
tiveswillonlyslightlydifferbetween children
withandwithoutrecurrences.
Previous studieshaveshownthat90%of
children's firstrecurrences willoccurwithin
twoyearsoftheinitialseizure.4 57Inthis
study35(97%)of36children hadtheirfirst
recurrence withintwoyears.Seventy percent
ofthechildren without recurrences were
followed formorethantwoyears.
Overall risksofoneandtworecurrences
inthisstudywere31%and15%andare
similartorecurrence risksinearlierstudies.3-7
Atwofoldincrease intheriskofonerecurrence
inchildren withafirstdegreerelativeaffected
byfebrileseizures wasalsofoundinother
studies.' 5-7
Inprevious prospective clinicbasedstudies
25%ofallchildren withafirstfebrileseizure
hadafirstdegreerelativeaffected byfebrile
Table3Recurrence risksinrelationtofebrileseizurehistoryofrelatives
Noatrisk Nowith Hazardratio
Feature (n=115) recurrence(risk*) (CI)t
Noofsiblingswithfebrileseizures
0 103 30(29) rc1 11 6(55) 2-4(10to58)
Noofparentswithfebrileseizures
0 101 28(28) rc
1 13 8(62) 3-1(1-4to6-7)
Noofgrandparents anduncles/aunts withfebrileseizures
0 96 31(32) rc31 19 5(26) 0-8(03to20)
Noofcousinswithfebrileseizures
0 102 31(30) rc¢1 13 5(38) 1-4(0-5to35)
*Kaplan-Meier estimates(0/%)oftwoyearcumulative incidence.tUnivariate hazardratioswith95%CIcompared withreference category(rc).seizures.6 21Inthisstudyaffected firstdegree
relatives werepresent in18%ofcases.The
percentage ofchildren withaffected first
degree relatives innon-West European
children (10%)wassignificantly lowerthan
thepercentage inWestEuropean children
(25%).Thereisnoreasontoassumealower
incidence offebrileseizures inMediterranean
orCaribbean children thaninWestEuropean
children. Morelikelytherehasbeenunder-
reporting ofaffected firstdegreerelatives,
possibly causedbyreluctance torevealthe
occurrence offebrileseizures totheinves-
tigatororbyhampered accesstoparentsliving
abroad. Thepresence ofaffected firstdegree
relatives yieldedsimilarrecurrence hazards in
WestEuropean andinnon-West European
children, whichisindicative ofnon-selective
underreporting ofrelatives affected byfebrile
seizures.
Nosignificant increase infebrileseizure
recurrence wasfoundinchildren withaffected
seconddegreerelatives orcousins only;the
historyoffebrileseizures inseconddegree
relatives andcousinstherefore appearstohave
littlevalueinestimating achild'sriskofrecur-
rence.
Thecrudeproportion ofaffectedrelatives –
thatis,theproportion offirstdegreerelatives
affected byfebrileseizures -yielded much
morediscrimination ofrecurrence risksof
febrileseizures thancommon familyhistoryof
febrileseizures. Withtheuseofthiscrude
proportion children withasixfoldincreased
riskofoneandfivefoldincreased riskoftwo
recurrences couldbeidentified -thatis,those
children ofwhom50%ormoreofthefirst
degree relatives wereaffected. Risksof
0L)
0L)
0
01).0
0~100
90
80
70
60
50
40
30
20
10
0
0051.0 1-5 2-0 2-5
Yearssincefirstseizure
Figure3Probability ofrecurrence inrelationtotheproportion ofrelativesaffectedbyfebrileseizures.398 on February 26, 2020 by guest. Protected by copyright. http://adc.bmj.com/ Arch Dis Child: first published as 10.1136/adc.70.5.395 on 1 May 1994. Downloaded from

Familyhistoryandrecurrence offebrileseizures
Table4Recurrences risksforfebrileseizuresaccording tothecrudeproportion andtheweightedproportion offirstdegree
relativesaffectedbyfebrileseizures
1Recurrence u2Recurrences
Hazardratio(CI)t Hazardratio(CI)t
NoatNo NoProportion risk(nisk*) Univariate Multivariate (nisk*) Univariate Multivanrate
Crudeproportion
0 9425(27)rc rc 13(14)rc rc0<proportion<0 5156(40)1-7(07to4-1)2-2(09to56) 2(13)
¢05 65(83)6-3(2-4to16-8)6-8(2-4to19-1) 3(50) 5-2 (1-5to18-3)5-7(1-4to18-5)
Weighted proportion
0 9425(27)rc rc 13(14)rc rc0<proportion<0 5135(38)1-7(0-6to4-4)2-2(0-8to6-1) 2(15)
¢05 86(75)4-4(1-8to10-8)5-1(2-0to12-8) 3(47)3-4(1-0to11-8)3-5(1-0to12-5)
All 11536(31) 18(16)
*Kaplan-Meier estimates (%)oftwoyearcumulative incidence.tHazard ratioswith95%CIcompared withreference category (rc).
tworecurrences inthesechildren werealso
significantly increased.
Although expected intheory, noimprove-
mentintheriskassessment oftherecurrence
offebrileseizures wasachieved byuseofthe
weighted proportion ofaffectedrelatives, with
adjustment fortheattained ageandsexofthe
relatives. Thismaybeduetothefactthattwo
thirds(227of348)offirstdegreerelatives were
parents. Thusmostrelatives werenolongerat
riskofafebrileseizureattheinitialseizureof
theindexchildren.
Bothgenetic andenvironmental mech-
anisms have beensuggested forthesuscepti-
bilitytoanincreased riskoftherecurrence
offebrileseizures.5 1522-25Inourstudy
recurrence riskforfebrileseizures inchildren
withaffected siblings weresimilar torecur-
rencerisksinchildren withaffected parents.
Thesefindings support amainly genetic
mechanism because environmental riskfactors
wouldhaveinduced alargereffectonthe
recurrence riskofsiblings thanofparents
affected byfebrileseizures. Anautosomal
dominant modeoftransmission canbe
assumed inchildren withproportion valuesof
0-5ormore.Richetalpostulated anautosomal
dominant modeofinheritance inchildren with
frequentrecurrences.26
Weconclude thatafirstdegreefamily
history isofmajorimportance inthe
assessment oftherecurrence riskforfebrile
seizures; second andthirddegreefamily
histories appeartobeofminorimportance.
Theproportion ofaffectedfirstdegreerelatives
yieldsthehighestdifferentiation oftherecur-
renceriskforfebrileseizures. Thisproportion
ofrelatives affected byfebrileseizures may
proveausefultooltoassesstherecurrence risk
offebrileseizures indailypaediatric practice
because ofthesimple assessment andthe
uncomplicated calculation oftheproportion.
Theauthors thankCMvanDuijn,PhDforherhelpful
comments.
1Offringa M,Hazebroek AAJM, Derksen-Lubsen G.
Prevalence offebrileseizures inDutchschoolchildren.
PaediatrPerinatol Epidemiol 1991;5:181-8.2Forsgren L,Sidenvall R,Blomquist HK,Heijbel J.A
prospective incidence studyoffebrileconvulsions. Acta
Paediatr Scand1990;79:550-7.
3VerityCM,ButlerNR,Golding J.Febrileconvulsions ina
national cohortfollowed upfrombirth.I:Prevalence and
recurrence inthefirstyearoflife.BMJ1985;290:
1307-10.
4NelsonK.B,Ellenberg JH.Prognosis inchildren withfebrile
seizures. Pediatrics 1978;61:720-7.
5Offringa M,Derksen-Lubsen G,Bossuyt PM,LubsenJ.
Seizure recurrence afterafirstfebrileseizure: amultivari-
ateapproach. DevMedChildNeurol1992;34:15-24.
6Knudsen FU.Recurrence riskafterfirstfebrileseizureand
affectofshorttermdiazepam prophylaxis. ArchDisChild
1985;60:1045-9.
7Annegers JF,Blakley SA,Hauser WA,Kurland LT.
Recurrence offebrileconvulsions inapopulation-based
cohort.Epilepsy Res1990;5:209-16.
8FarwellJR,LeeYJ,HirtzDG,Sulzbacher SI,Ellenberg JH,
NelsonKB.Phenobarbital forfebrileseizures -effectsonintelligence andonseizure recurrence. NEnglJfMed
1990;332:364-9.
9Chessare JB,Berwick DM.Variation inclinicalpractice in
themanagement offebrileseizures. PediatrEmergCare
1985;1:19-21.
10Knudsen FU.Effective short-term prophylaxis infebrile
convulsions. JPediatr 1985;106:487-90.
11AutretE,BillardC,Bertrand P,MotteJ,Pouplard F,JonvilleAlP.Double-blind, randomized trialofdiazepam
versusplacebo forprevention ofrecurrence offebrile
seizures. JPediatr1990;117:490-4.
12RosmanNP,ColtonT,LabazzoJ,etal.Acontrolled trialofdiazepam administered duringfebrileillnesses toprevent
recurrence offebrileseizures. NEnglJMed1993;329:
79-84.
13ElRhadiAS,BanajehS.Effectoffeveronrecurrence rateof
febrileconvulsions. ArchDisChild1989;64:869-70.
14WolfSM,CarrA,DavisDC,etal.Thevalueofpheno-barbitalinthechildwhohashadasinglefebrileseizure: a
controlled prospective study.Pediatrics 1977;59:378-85.
15Frantzen E,Lennox-Burchthal M,Nyaard A,etal.Ageneticstudyoffebrileconvulsions. Neurology 1970;20:
909-17.
16Bethune P,GordonK,DooleyJ,etal.Whichchildwillhave
afebrileseizure? AmJDisChild1993;47:35-9.
17Millichap JG.Thedefinition offebrileseizures. In:Febrile
seizures. NewYork:RavenPress,1981:2.
18KaplanEL,MeierP.Nonparametric estimates fromincom-pleteobservations. Journal oftheAmerican Statistical
Association 1958;53:457-81.
19CoxDR.Regression modelsandlife-tables. Journalofthe
RoyalStatistical Society[B]1972;34:187-220.
20TausigTE,EllenMSW,Freeman W.Thenextbestthingtobeingthere:conducting theclinicalresearch interview bytelephone. AmJOrthopsychiatry 1988;58:418-27.
21BergAT,ShinnarS,HauserWA,etal.Aprospective study
ofrecurrent febrileseizures. NEnglJfMed1992;16:
1122-7.
22Baraitser M.Relevance ofafamilyhistoryofseizures. Arch
DisChild1983;58:404-5.
23Tsuboi T.Genetic aspects offebrileconvulsions. Hum
Genet1977;38:169-73.
24HauserWA,Annegers JF,Anderson VE,Kurland LT.The
riskofseizuredisorders amongrelatives withfebrilecon-
vulsions. Neurology 1985;35:1268-73.
25Forsgren L,Sidenvall R,Blomquist HK,Heijbel J,Nystrom L.Anincident case-referent studyoffebrilecon-
vulsions inchildren: genetical andsocialaspects.Neuropediatrics 1989;21:153-9.
26RichSS,Annegers JF,HauserWA,Anderson VE.Complexsegregation analysis offebrileconvulsions. AmHumGenet1987;41:249-57.399 on February 26, 2020 by guest. Protected by copyright. http://adc.bmj.com/ Arch Dis Child: first published as 10.1136/adc.70.5.395 on 1 May 1994. Downloaded from