Desk Reference 1997). Conversely. other antiepileptic ... Physician's Desk. Reference 1997) and ... Seven children had C'I' or MRI scans which. \\'ere normal.
Inappropriate use of carbamazepine and vigabatrin in typical absence seizures A P J Parker* MKCP; A Agathonikou MD: R 0 Robinson FKCP; C P Panayiotopoulos MD PhD FRCP; Department o f Clinical Neurophysiology and Epilepsies. St Thomas' I Iospital and Department o f I'aetliatric Neurology, Guy's Hospital. I.ondon, UK. "(.'~rr~spo,ir(etice tofirst a i r t h ((1 Department o f Clinical Neurophysiolog and Epilepsies. St Thomas' I lospital. I.ontlon SEI 7131, UK.
Carbamazepine and vigabatrin are contraindicated in typical absence seizures. Of 18 consecutive referrals of children with resistant typical absences only, eight were erroneously treated with carbaniazepine either as monotherapy or as an add-on. Vigabatrin was also used in the treatment of two children. Frequency of absences increased in four children treated with carbamazepine and two of these developed myoclonic jerks, which resolved on withdrawal of carbamazepine. Absences were aggravated in both cases where vigabatrin was added on to concurrent treatment. Optimal control of the absences was acheved with sodium valproate, lamotrigine, or ethosuximide alone or in combination.
Typical absences (TA) are epileptic seizures manifested mainly with impairment of consciousness and 2.5 t o 4 Hz Senemlizetl spike and slow-wave discharges. They are casily induced by hypcmentilation. l'hcymaybe the onlysehrr: type, ;LZ inchildhood absence epilepsy (pyknolepsy), or combined with other seizures, as in juvenile absence or juvenile myoclonic epilepsy (British National Formuhy 1997,Rnayiotopoulos 1997). T A usually respond well to the appropriate antiahsence (Irugs such ;isethosuximide and sodium valproate alone o r in combination (British National Formulary 1997. Physician's Desk Reference 1997). Conversely. other antiepileptic drugs such ascarbamazepine (Cereghhinoet al. 1974. Hornet al. 1986, Joyce 1992, British National I:ormulaty 1997. Physician's Desk Reference 1997) and vigabatrin (Ferrie and Kohinson 1995) make them more frequent. As children are still referred to us that have been trtxted with these drugs. we report the inappropriate management o f typical absence seizures and indicate correct treatment strategies.
Method W'e retrospectivelyexamined our records of 1003 to 1905, to identify children referred t o our epilepsy clinics with TA only. All have unequivocal clinical and EEC,hidec+EIiG documentation ofTA. Children who also had tonic-clonic, myoclonic. o r other types of generalized seizures were excluded. Kcsponse t o each treatment \\'asobserved by t h e parents and reported t o the clinician as an increrLse. decrease. or no change in thc frcquency of absences. and \KLS rccortletl in the hospital notes. 'I'ypical absences are detined as: 'generalized epileptic seizures clinically characterized by impairment of consciousness only (simple absences) or impairment ofconsciousncss combined with mild clonic. atonic. tonic. and autonomic components, and automatisms (complex absences). The ictal-EEG signature of an absence attack is a hilatcral usually regular and symmetrical discharge of3 Hz (2.5 t o 4 Iiz) spike and slo\v-\vave cchplexes.' (Panayiotopoulos 1997). Results Fourteen girls and four boys with 1 3 as the onlv seizure type were idcntified. Mean age at seizure onset was 5.2 years (range 4 to 12 years in all but one. in whom onset WLS at 12 months). EEG, videcrEEG. o r both confirmed the diagnosis in all children. Seven children had C'I' or MRI scans which \\'ere normal. lnitial treatment w s w i t h sodium valproate in 16 children and cthosuxiniide i n the other nvo. Eight children had been treated with carbamazepine either as monothcrapy (five children) or add-on (l:dde I). Carbamazepine was introduced clue to continuation of absences in seven children and weight gain in the other. Frequency of absences increased in four children and did not change in two.'rww children developed additional myoclonic jerks with carbamazepine. which resolved on withdrawal. Carbamazepine was discontinued due to a skin rash in two others. At the time o f carbaniaxpine initiation, the EEG reports were o f TA with generalized discharges in all except patients 3 and 6 (see Table 1) where the discharges of the 'li\ were described as 'bilated' with 'some asymmetrical bursts' in one and 'some temporal sharp waves' in the other (see Fig. 1 for subsequent EEG recording of this child). Two children also received vigabatrin with a dramatic
increase of seizures which prompted hospital admission in o n e (see illustrative case history below). Carhamazepine or vigabatrin were stopped in all children; absence frequency decreased o n treatment with sotliu m valproate. lamotrigine, and ethosuximide alone or in Combination. I1.I.I' S'IHATIVt CASE f 1 l5TOlW
This girl (patient 6) was seizurc-frec until aged 6 years when s h e developed typical abscnces (Fig. 1). Initial treatment with sodium valproate ( u p t o 34 m g k g ) had only a niinor beneficial effect. Absences continued with occ;isional urine incontinence. Ethosuximide at 8mgkg was added o n with n o bcnetit, a n d was replaced by carlmnazcpine, phenytoin, a n d vigabatrin consecutively (see 'lible I). Absences a n d
learning difficulties increased with carbamazepine and vigabatrin. O n combining carbamazepine with vigabatrin a n d sodium valproatc she worsened dmmatically, requiring hospital admission. Carbarnaxpine and vigabatrin were withdrawn. O n introduction of lamotrigine the absence frequency decreased and she remains well controlled with less than o n e absence per day o n sodium valproate (24 mgkg) and lamotrigine ( 4 m g k g ) for the last 4 years,.
Discussion in o u r report w e have relied o n parental assessment of absence frequency Absolute absence frequency is not possible to ascertain in a retrospective analysis and even in prospective studies t h c majority of repons rely o n parental assessment of trend rather than numbers of abscnces.
Table I: Age at onset of absences, and treatment regimes and responses (chronologicalorder) of children treated with carbamazepine
No change Erhosuximidc \Veight gain lthosuxiniide+sodium v;ilproate 1ncre;isc in hF Ethosuximitle t cirbiniazepinc I)ecrc;rw in h F firhosusimid~+sodiumvalproatc Ikcreasc in AF fit hosuximidc+sotfium valproate + lamotrigine Sodium valproatc+ lamotriginc Ccssition of abscnces No change 7 Sodium valproatc Hash Carhamaxpine iithosuximitle Cessation olabsenccs No change 1 Sodium valproatc No change Carbamazepine N o changc Sodium valproate Decrease in AF Sodium vnlproare+lamorriKinc No change 5 Sodium valproate No change Carhamaxpine No change Phenytoin Decrease in AF Phcnytoin+cthosuxiniide Decrease in AI; iir hosuximidc Decrease in AF lthosuximidc + lamorrigine Nochange 6 Sodium valproare Incrcasc in A F Carhamazepine Increase in AF Carbamazcpinc+vigabatrin Sodium valproate Decrease in AF Decrease in AF Sodium valproate + lamotriginr Decrease in A F Sodium valproatc 5 N o change Sodium valproate +ethosuximide Increase in AF Sodium valproak +carbamazepine No change Sodium valproatc + phenyroin Increase in AF Sodium valproate+vigabatrin Increase in AF Sodium valproate + vigahatrin + carhamaxpine Dccreasc in AF Sodium valproate+ lamotrigine Ikcreasc in AP 5 Ethosuximide N o change Ethosuximldc + sodium vilprolte Ethosuximidc +carbarnaxpine Rash I)ccrcase in AF Sodium valproaie Decreasc in N: Sodium valproatc + rthosuximidc 7 N o change Sodium valproate Ethosuximide Decrease in AF Carbamazepine Increase in AF Carbamazepine+ethosuximide Decrease in AF Sodium valproate Decrease in AF
5
1
7
3
5
6
7
8
AF, absence frequency.
5 18 Devehpmetital Medicine & Child Neumlogv 1998,40: 5 17-5 19
It has been known for more than 20 years that carbamazepine is not indicated in the treatment o f absence seizures (Cereghinoet al. 1974),a n d this isclearlystated in all relevant t c x t h o o b , t h e British National Rtrrnulary, a n d the Physician's Desk Reference. Furthermore, carbamazepine may exaggerate or induce typical o r atypical absences (Horn et al. 1986,Joyce 1992). In this study sodium valproate or ethosuximide. o r both together, were appropriately used as the first-line drugs. However, w h e n these failed ti) s t o p or iniprove t h e absences, 44% of the children were subsequently treated with carbamazepine. It is also of concern that in four of the eight child r e n c;lrbam;acpine \\*as used as a second-line drug to stxlium valpwate or ethosuximide. We were unable t o find a n expl;ination for t h e use of carbarnaxpine and vigabatrin in thcse children with 1A which were unequivocally contirnied o n EEG. An inllucntinl factor may have been the trend to treat 'epilepsy' in a unitary fashion either with sodium valproate or carbamazepine, neglecting seizure and syndrome characteristics. l h a t the TA were resistant to sodium valproate o r ethsuximide may be another misleading factor. but in these cases even if adequate doses of sodium valproate, ethosuximide, or both. failed, there are alternatives. Small doses of laniotrigine added to sodium valproate controlled similarly resistant C;LSCS ofabsences seizures (Ferrie e t al. 1995). It is not sufticiently appreciated that automatisms and xsymmetric myoclonic jerks frequently accompany TA a n d often result in them bcing described as atypical a n d misinterpreted as partial seizures. The term atypical is often misapplied to both t h e clinical features and EEG characteristics a n d does not necessarily imply that t h e seizure is partial. In addition, the use of 'bilateral' rather than 'gencralized' t o describe the 3 t o 1 I-lz generalized discharges a n d reports o f sharp \vaves or minor asymmetries, which occurred in two subjects in our repon. may lead clinicians to think they are
dealing with partial seizures with secondary generalization. Absences are pharmacologically different from d1 other types of seizures and are mediated mainly,by gamma-aminobutyic a~citl,~ (tiABt%d activation (Duncan and Panayiotopoulm 1995). GABA mimetics such as vigabatrin or ngonists such zs baclofen make them worse \vhile (;MA antagonists produce a dose-dependent abolition of absences in animal models (Duncan and Panayiotopoulos 1995. Hosford and Nhng 1997). Vigabatrin, a n irreversible inhibitor of the degndative enzyme GABA transamin;w, also exacerbates absence seizures in humans (Ferric and Kobinson 1995. h n a y i o t o p o u b s 1997). 'I'hat vigibatrin is contraindicated in ccnain forms o f myoclor$c epilepsies has been emphasized (British National 1:ormul;uy 1997). Its harniful effect on children with absences hzs not received the s m e attention (Appleton 19%. British National I~ormu1;u). 1997). thus it may increasingly becomc more o f 3 hazard xs indicated by the case histories of patients j and 6 of this report and by other reports (iimic and Kobinson 1995, Panayiotopoulos et al. 1997). lifficacioiis d r u g s used in t h e treatment of absence seizures include sodium valproate. ethosuximide, or o t h e r succinimides, lamotrigine. clonazepam. or o t h e r henzodi- a x p i n e s a n d acctazolamide (Duncan a n d Panayiotopoulos 1995. Panayiotopoulos et al. 1997).
Conclusion C a r b a m z e p i n e and vigabatrin should not be prescribed for t h e treatment ofTA seizures.
References
hpplcron RE. ( 1396) N e w antiepileptic drugs,Arc/Jiwsojllismsr in CiJiIdhod 75: 256-62. British iVuiit~~riil~~rrrrtilor).. .$it/) c h . ( 1997) Iundon: British
Medical Association. CcrediinoJJ. BrcxkJT, van hletcrJC, PenryJK. Smith LD. \Y'hite RG. (177.4)Carbamazepine for epilepsy: a controlled prospcctivc evaluation. rVertro1o.g 24: 4 0 1- 10. Duncan JS. Panayiotopoulos CP. editors.( 1995) 7jpicufAhserrces nftd Relared Syrtdrotrres. London: Churchill Communications. Ferric CI). Robinson RO. Knott C. Panayiotopoulos CY (19%) Ianotrigine as an add on drug in typical absences. Acts iVcirrologicG Sccoiditraaicu91: 200-2. _ - (1995)l h e clinical efficacy ofvigabatrin in childwn. Reideri~of
Fp2 Fa
Crmtenipporur).fbanrrncotherupp 4: 469-76. H o r n S. Ater SU.Hurst DI.. (1986)Carbamazclpineexacerbated
epilepsy in children and adolescents. P~~~iutric?\l~.rtrolo~)i 2: 3 4 0 4 .
p3 O1
~
200IJv
r
L
~
~
P
~
f
v
~
p
'
,
\
IS
Figure 1: EEG ofpatient 6 at uge IOyrars clemonstrcitingci tj pica1 absence ivbi/e tJje child counted. TJie EEG discburge consistsofgeneralized spike and occusiotiul~~ po~spike sfow-ivaveat.3Hzsloiviri~doivti to2.5 Hxat tbeenrl. Cfitiiccrf& she stoppedcoitntitig atid opened her e-yes rvitliit; thejirst 2 sfrom the omet of the EEG dischrge, there wus rhythmic niyoclonris of the eyebroivs,her eyes trirned up and to the right. hlarkedperiorul and bund airtoniutisnis, arid cheruitig niooements occicrreri in the middle of the discharge.
Hosford DA, WngY (1997) Utility ofthe Iethugic (Ihtlh) mouse model of absence seizures in prcdictingthe effects of ~ lamotrigine, , * ~ vigabatrin, ~ - tigabine. gabipentin. and topinmate against human abscncc seizures. Epilepsia 38: 408-14. Joyce C. (1972) Carbamazepine is not for petir ma1seizures. /lrrrericatif&unrmcyNS32:6. Panayiotopoulos CE (1997)Absence Epilcpsics. In: Engel JJ. Pedley IA, editors. Epilepsy: A Cof?rpre/Jefrsiw Texlbook. Philadelphia: 1.ippincott-Raven.p 232746. -t\gathonikou A, Sharoqi LA. ParkerhPJ. (1997) Vigabatrin '. aggravates absences and absence status. iSeitrolog).49: li67. Phpicinris'Desk Referetice. 5lst eilrr. (1937) hfonrvalr. NJ: Medical liconomics.
Carbamazepine and Vigabatrin inlypicd Absence Seizures k y f u r k c r
el a!. 519
