554 Archives ofDisease in Childhood 1991; 66: 554-557

PERSONAL PRACTICE Arch Dis Child: first published as 10.1136/adc.66.4.554 on 1 April 1991. Downloaded from

When does an EEG contribute to the management of febrile ?

Gregory Stores

The above question implies that the elec- group. However Livingstone's neat distinction troencephalogram (EEG) can be of some value between simple and atypical febrile seizures has in the care of children with febrile seizures. Not also been challenged, again perhaps particularly everyone would agree. For example Aicardi sta- by Aicardi who asserts that although a large tes uncompromisingly that 'EEG tracings are majority of simple febrile seizures are brief, useless as they do not help predict the occurr- bilateral and benign in outcome, some chil- ence of later '.' Some accounts are so dren have a combination of simple febrile sei- disdainful that they do not mention electro- zures and long lasting unilateral seizures in spite encephalography at all.2 However other authors of a strong family history of febrile seizures.' are less sure: 'The role of the electroencephalo- Chevrie and Aicardi point out that even some gram in the work-up of febrile seizures remains febrile seizures with residual neurological signs controversial'.' are not necessarily followed by epilepsy.7 In The possible value of the EEG in these cir- addition the severity of the attacks may not be a cumstances depends to some extent on the good predictor of outcome as almost 90% of definition of febrile seizures and the possible patients who develop epilepsy have had only identification of different types. Both are still brief bilateral seizures.8 In a later publication contentious issues. Aicardi and Chevrie concluded that unilateral, Regarding definition, the National Institutes prolonged febrile seizures are a heterogeneous of Health (NIH) consensus statement excluded group containing some children who have sei- seizures with evidence of intracranial infection zures because of genetic vulnerability to fever or a definite cause, and also children with pre- but also others who convulse because of vious non-febrile seizures.3 Although these cri- acquired injury.9 Aicardi concludes: 'sim- teria are accepted in some child texts ple seizures-defined as brief, bilateral attacks http://adc.bmj.com/ (for example, Menkes4), other authors point to in previously normal children-are not neces- the difficulties of achieving a strict definition in sarily true febrile , in the sense of these terms, for example Brett.' Aicardi in par- seizures entirely resulting from an abnormal ticular is critical of the NIH consensus defini- constitutional response to fever.' Conversely, tion: 'The absence of intracranial infection and complex febrile convulsions or febrile statuses of recognised acute neurological illness remains can be true febrile convulsions, . . . the com- an unproved assumption in at least some cases'.' plete separation between simple and atypical on September 29, 2021 by guest. Protected copyright. As for subgroups, there is general (though febrile convulsions reported by Livingstone has not complete) agreement with the distinction not been confirmed in any other series'. This made by Livingstone between 'simple febrile view should not be construed as excluding the convulsions' and 'epileptic seizures precipitated possibility of defining in the future convincing by fever' or 'atypical febrile convulsions' which subgroups of children with febrile seizures. are also sometimes called 'complex' or 'com- Indeed, for example, Aicardi himself points out plicated' febrile seizures.6 In Livingstone's that febrile seizures of early onset and long account simple febrile seizures, accounting for duration with only a low grade fever and a posi- 97% of the total, were brief generalised seizures tive response to photic stimulation on elec- in children with a family history of febrile sei- troencephalography can be the first manifesta- zures, in contrast to atypical febrile seizures that tion of 'severe of infants'". were usually prolonged and often focal or lat- eralised. Only 3% of the former group develo- ped later afebrile seizures compared with 93% Reports of EEG abnormalities in children University of Oxford of the latter group. In more recent years it has with febrile seizures (table) Department of been claimed that this poor prognosis group is (1) Ictal EEGs Psychiatry and Park further characterised by such features as neuro- Recordings during acute febrile illness with sei- Hospital for Children, Oxford logical or developmental abnormality before the zures have rarely been reported. They can be of 1 as in Correspondence to: onset of seizures, onset under the age year, expected to show generalised rapid spiking Dr Gregory Stores, and frequent or prolonged febrile seizures. any tonic or tonic-clonic but according Section of Child and Adolescent Psychiatry, These descriptions begin to suggest a possible to Lennox-Buchtal some show severe lateralised Park Hospital for Children, role for in helping to electroencephalographic changes of this type." Old Road, Headington, identify the nature of underlying cerebral In her study, however, no distinction was made Oxford OX3 7TQ. pathology in children assigned to the atypical between simple and atypical febrile seizures. When does an EEG contribute to the management offebrile seizues? 555

EEG abnormalities in children with a history of febrie seizures. (b) Generalised spike wave discharge seizures occurring in the resting EEG or in response to Ictal (rarely reported) Generalised spiking overbreathing. This pattern was seen in 49% Lateralised spike wave discharge with age of onset usually between 3 and 6 years. Arch Dis Child: first published as 10.1136/adc.66.4.554 on 1 April 1991. Downloaded from Postictal Slow activity (c) Generalised spike wave in response to photic Spike wave or spikes Serial EEGs Bisynchronous theta activity stimulation was seen in 42% mainly in children Bisynchronous spike wave at rest under the age of 5 years. and during overbreathing Bisynchronous spike wave on photic From this and other reports it appears that stimulation generalised spike wave discharge is unusual at Focal spikes or sharp waves the younger age for febrile seizures (that is, Hypnagogic paroxysmal spike wave under 2 years) appearing on average about six- teen months after the first up to as late as four years after it. 11 20 The discharges can be intermittent with sometimes months or (2) Postictal EEGs years apparently in abeyance, although the sam- (a) Recordings taken on the same day as the sei- pling error of repeated brief electroencephalo- zure have been reported to contain slow activ- graphic recordings has to be remembered. ity in as many of 88% of cases. The same type of Focal spike or sharp wave discharges have abnormality is described in about a third ofchil- been reported in between 8% and 10% espe- dren between three and seven days after a cially where a persistent slow have focus was seizure.'2 This slow activity is predominantly seen previously usually after a prolonged posterior in distribution and, although mostly seizure. 1 However, such discharges are also seen bilateral, is often asymmetrical or even unilat- after brief febrile convulsions, looking very eral. It tends to resolve over about a week. The much like centrotemporal spikes of later reports by Pampiglione'3 and Wallace and childhood.2' Kajitani et al reported this finding Zealley'4 suggested that such abnormalities can in 4-2% of children on follow up.17 be expected in association with viral illnesses Another obvious abnormality in the later whether or not seizures occur, although not in EEGs of children with febrile seizures is that relation to changes of temperature.'" However described by Alvarez et al.22 They emphasised Rantala et al found no difference in the early the occurrence of generalised, paroxysmal, high electroencephalographic findings (including the amplitude rhythmic 3-5 cycles per second slow presence of abnormal slow activity as well as activity with intermixed spikes in drowsiness spikes or sharp waves) between those with and ('hypnagogic paroxysmal spike wave activity' or those without a proved viral infection.'6 'hypnagogic PSW') in perhaps a quarter of chil- dren with febrile seizures, mainly in the second (b) Spikes or spike and wave discharges, local- year of life and rarely younger than this. This ised or generalised, are rare in the early postictal feature was not seen in a group of normal con- period. Frantzen et al gave a figure of 1-4%12 trols and appeared in only 9% of a group ofchil- and Kajitani et al suggest 3%."' Both these dren referred for electroencephalography for http://adc.bmj.com/ figures are close to the 190/o given by Eeg- neurological problems but without febrile sei- Olofsson'8 as the incidence of focal sharp activ- zures. There appeared to be no clinical differ- ity in normal young children. ence between children in the febrile seizure Rantala et al suggest that the presence of any group with or without the EEG abnormality. of these abnormalities does not distinguish between simple febrile seizures (defined in their

study as lasting less than 15 minutes and with- Prospective studies on September 29, 2021 by guest. Protected copyright. out focal features) and complicated febrile sei- These various abnormalities are very interesting zures (that is, lasting longer than 15 minutes, but there seems to be universal agreement that with focal features or recurrence within 24 none of them bears any convincing association hours).'6 In this investigation febrile seizures with the eventual development of epilepsy (that were defined as a first seizure associated with is, recurrent afebrile seizures of primarily cere- fever irrespective of age and excluding children bral origin) in these children. Some key studies with earlier non-febrile seizures, evidence of a of febrile convulsions concerned with risk brain infection or metabolic encephalopathy. factors for later afebrile seizures did not speci- fically assess the role of electroencephalo- graphy.23 24 However the discussion in the NIH (3) Serial EEGs over several years consensus statement on febrile seizures is Frantzen et al reported that 290/o of children instructive on this point.2' The discussion showed some form of 'epileptic' activity in the started with the assertion that on general princi- few years after their last febrile seizure.'2 Doose ples electroencephalography findings ought to et al carried out a longer electroencephalog- be predictive of later epilepsy and therefore raphic follow up study on children with febrile should form one of the considerations on which seizures to the age of 11 to 13 years and reported each child's management should be based. that 81% had abnormal EEG patterns. However, a series of studies were than quoted These authors described three types ofabnor- by Hauser (USA), Thorn (Denmark), and mality all considered to represent genetically Camfield (Canada) all of which indicated no determined patterns: (a) bilaterally synchronous association between early electroencephalo- and regular 4-7 cycles per second (theta) graphy findings and later seizures. Incidentally, rhythms seen mainly in parietal regions occur- these general conclusions seem to apply to boys ring in 54% of his series of children with febrile and girls equally. 556 Stores

Conclusions concerning the clinical value of management of children with febrile seizures, EEG investigations in children with febrile the high rate of the various EEG abnormalities seizures mentioned earlier (including the different forms Clinical value of seizure activity) on long term serial elec- Arch Dis Child: first published as 10.1136/adc.66.4.554 on 1 April 1991. Downloaded from An early postictal standard EEG will not be troencephalographic follow up of children with helpful in the following respects: (a) it will not febrile seizures, means that there is a serious distinguish between clinically simple and atypi- risk of misdiagnosing epilepsy on the basis of cal seizures and (b) it will not particularly help such EEG findings alone. Seizure discharge in the identification of a cerebral infective found incidentally in older children should not aetiology. (c) EEG findings lack predictive necessarily be construed as evidence of current value for the later occurrence of either further epilepsy. Especially in the presence of a history febrile or later afebrile seizures. As Aicardi says, of febrile seizures, such abnormalities may well epileptic activity will be found in a high propor- be simply evidence of a persistent genetically tion of older children who are less likely to determined electroencephalographic trait un- develop recurrent seizures and later epilepsy, associated with the risk of afebrile seizures. whereas they will almost never appear in the EEGs of young infants that are at maximum risk Research of recurrence, later afebrile seizures, and severe Another possible although very different place febrile seizures.' It is these discrepancies that for electroencephalographic investigations in lead Camfield to say in the above NIH discus- children with febrile seizures concerns research, sion that 'It makes more trouble than it is especially the investigation of genetic factors. worth' to carry out recordings in these circum- The morphological similarity mentioned earlier stances; 'you wish you had not done the EEG', between some of the focal spike or spike wave because with an abnormal result discussions discharges seen in some children years after the with parents about its significance can be very febrile seizures and the centrotemporal spike difficult. For these reasons electroencephalogra- discharges in benign rolandic epilepsy of child- phy cannot be recommended as a routine mea- hood has led to the suggestion that there may be sure in children with a febrile seizure. some genetic connection between febrile sei- Perhaps, however, there is a limited place for zures and this form of epilepsy.'7 Similarly electroencephalographic studies in connection Doose et al have suggested that various geneti- with febrile seizures in those cases where there cally determined EEG patterns can be found in is clinical reason to suspect cerebral pathology. a high proportion of patients with febrile sei- If the child showed developmental delay or zures, that is, biparietal theta rhythms, spike other evidence of neurological impairment wave activity in the resting state and on over- before the occurrence of the febrile seizures, or breathing, and that provoked by photosensitiv- if the first seizure occurs below the age of 12 ity as described earlier. 9 Alvarez et al suggested months, or if the seizures contain partial (focal) that hypnagogic paroxysmal spike wave is also a

features in the ictal or postictal phase, it seems genetically transmitted EEG pattern.22 http://adc.bmj.com/ prudent to consider the possibility that the Genetic studies are not the only type of child's seizures may be evidence of brain pathol- research needed in the field of febrile seizures. ogy, perhaps structural. From the above The possibility of defining clinically important account it is clear that these clinical features do subgroups was mentioned earlier in connection not necessarily suggest a brain lesion and in with severe myoclonic epilepsy of infants. most cases the findings on further investigation Whatever aspect of febrile seizures is resear- will be normal, but in some instances they will ched, as Alvarez and his colleagues rightly point be found to indicate a significant brain abnor- out,22 published studies contain clear inconsis- on September 29, 2021 by guest. Protected copyright. mality. tencies not only in the criteria used to compare In addition, if a child has suffered a pro- patient groups in different studies of febrile sei- longed febrile seizure, especially if it is followed zures, but in the way electroencephalographic by residual neurological signs or developmental information has been obtained. They are right regression, the possibility of structural change to recommend, certainly for research purposes, caused by the convulsions needs to be consi- that recordings should be concerned with natu- dered and investigated in a similar way in view ral sleep and wakefulness, drowsiness, at least of the evidence that various degrees of brain stages 1 and 2 of non-rapid eye movement sleep, damage can be sustained during the course of a arousal, and wakefulness. Failure to standardise prolonged febrile seizure.9 the procedure makes comparisons between dif- Electroencephalography is not the investiga- ferent studies very difficult. Certainly if this tion of choice in considering structural pathol- sequence is not followed there will be a failure ogy and computed tomography or magnetic to demonstrate not only hypnagogic paroxysmal resonance imaging would be more appropriate. spike wave but other spike wave discharges dur- It can be a useful ancillary investigation capable ing sleep with which hypnagogic paroxysmal of suggesting persistent brain pathology (even spike wave might be confused. in the presence of a normal computed tomo- Further exploration of these possibilites gram) if it demonstrates an abnormality (usually dependent on electroencephalographic studies irregular slow wave disturbance with or without would be interesting and justified within the associated spike or sharp wave discharges) confines of a specific research protocol. Outside which is persistent. Clearly this implies the use the circumscribed clinical and research settings of serial EEG recordings perhaps over a period just mentioned electroencephalographic investi- of weeks rather than a single recording. gations of children with febrile seizures cannot Although not immediately relevant to the be recommended. When does an EEG contribute to the management offebrile seizures? 557

1 Aicardi J. The medical management of neonatal and infantile 14 Wallace SJ, Zealley H. Neurological, electroencephalo- seizures and of febrile seizures. In: Porter RJ, Morseili PL, graphic and virological findings in febrile children. Arch eds; The . London: Butterworth, 1986:206-26. Dis Child 1970;45:611-23. 2 Anonymous. The management of febrile convulsions. Drug 15 Minchom PE, Wallace S. Febrile convulsions: electroence- Ther Bull 1987;25:9-11. phalographic changes related to rectal temperature. Arch Arch Dis Child: first published as 10.1136/adc.66.4.554 on 1 April 1991. Downloaded from 3 Kendig EL, Dyken PR, Hernandez N, et al. Consensus Dis Child 1984;59:371-3. statement on febrile seizures. In: Nelson KB, Elienberg 16 Rantala H, Uhari M, Tuokko H. Viral infections and JH, eds. Febrik seizures. New York: Raven Press, 1981: recurrences of febrile convulsions. J Pediatr 1990;116: 301-6. 95-159. 4 Menkes JH. Textbook of child neurolog. Philadelphia: Lea and Febiger, 1985. 17 Kajitani T, Ueoka K, Nakamura M, Kumanomidou Y. 5 Brett EM. Paediatric neurology. Edinburgh: Churchill Living- Febrile convulsions and rolandic discharges. Brain Dev stone, 1983. 1981;3:351-9. 6 Livingstone S. Infantile febrile convulsions. Dev Med Child 18 Eeg-Olofsson 0. The development of the electroencephalo- Neuol 1968;10:374-6. gram in normal children from the age of 1 through 15 7 Chevrie JJ, Aicardi J. Duration and lateralisation of febrile years: paroxysmal activity. Newropaediatrie 1971;2: convulsions. Etiological factors. Epiepsia 1975;16:781-9. 375-404. 8 Ellenberg JH, Nelson KB. The efficiency of published 19 Doose H, Ritter K, Volzke E. EEG longitudinal studies in recommendations for the treatment of febrile seizures. In: febrile convulsions. Neuropediatics 1983;14:81-7. Nelson KB, Ellenberg JH, eds. Febrile seiwures. New York: 20 Tsuboi T, Endo S. Febrile convulsions followed by non- Raven Press, 1981:97-102. febrile convulsions. A clinical, electroencephalographic and 9 Aicardi J, Chevrie JJ. Consequences of in follow up study. Neuropediatrics 1977;8:209-33. infants and children. In: Delgado-Escueta AV, Wasterlain 21 Heijbel J, Blom S, Bergfors PG. Benign epilepsy of children CG, Treiman DM, Porter RJ, eds. Advances in neurolo. with centrotemporal EEG foci. A study of incidence in Vol 34. Status epilpticw. New York: Raven Press, 1983: outpatient care. Epiepsia 1975;16:657-4. 115-25. 10 Dravet C, Roger J, Bureau M. Severe myoclonic epilepsy of 22 Alvarez N, Lombroso CT, Medina C, Cantdon B. Paroxysmal infants. In: Roger J, Dravet C, Bureau M, Dreifuss FE, spike and wave activity in drowsiness in young children: its Wolf P, eds. Epileptic syndromes in in.fancy, childhood and relationship to febrile convulsions. Ekctroencephalogr Clin adolescence. London: Libbey, 1985:5847. Neurophysiol 1983;56:406-13. 11 Lennox-Buchtal MA. Febrile convulsions. A reappraisal. 23 Nelson KB, Ellenberg JH. Prognosis in children with febrile Electroencephalogr Clin Neurophysiol 1973;32 (suppl):1-32. seizures. Pediatrics 1978;61:720-7. 12 Frantzen E, Lennox-Buchtal MA, Nygaard A. Longitudinal 24 Annegers JF, Hauser WA, Elveback LR, Kurland LT. The EEG and clinical study ofchildren with febrile convulsions. risk of epilepsy following febrile convulsions. Neurology Elecsroencephalogr Clin Neurophysiol 1968;24:197-212. 1979;29:297-303. 13 Pampiglione G. Prodromal phase of measles: some neuro- 25 Nelson KB, Ellenberg JH, eds. Febrik seiures. New York: physiological studies. BMJ 1964ii:1296-300. Raven Press, 1981:297-300. http://adc.bmj.com/ on September 29, 2021 by guest. Protected copyright.