Clinical Review - 11 oct 1997

Clinical review

Fortnightly review

Epilepsy in childhood

B G R Neville

Childhood epilepsy requires integrated medical,

educational, and community services, and its treatment

spans acute and disability medicine.

There are many childhood epilepsies, and seizures

are the commonest paediatric neurological symptom.

Epilepsy

—

that is, susceptibility to continuing seizures

—

occurs in 0.5-1.0% of the population and is intractable

to current antiepileptic drug treatment in 20-25%. Epi-

leptic seizures, including febrile convulsions, occur in

3-5% of children. Epilepsy starts in childhood in 60%

of cases, and most of the clinically significant aspects of

the disease occur during childhood.

Method

I chose the studies on which I have based my review

because they support the development of clinical

guidelines. The six crucial clinical issues are:

x Identifying non-epileptic seizure syndromes

x Identifying benign epilepsy syndromes with a com-

paratively good outcome

x Identifying malignant syndromes in which continu-

ing and increasing disabilities occur and need detailed

investigation, aggressive treatment, and research

x Identifying, assessing, and managing the four major

disabilities that result from early brain damage or dys-

function and commonly coexist

—

epilepsy, motor

disorder (cerebral palsies), cognitive disability, and psy-

chiatric disability

—

particularly for the many epilepsy

syndromes that start in the first two years of life

x Ensuring that antiepileptic drugs are used in an

ordered fashion with audit of results

x Anticipating and managing status epilepticus.

Non-epileptic attacks

Non-epileptic attacks include simple blue breath hold-

ing attacks, white reflex asystolic attacks, reflex syncope,

benign paroxysmal vertigo, benign myoclonus of

infancy, vasovagal attacks, sleep phenomena, pseudo-

seizures, and Munchhausen syndrome by proxy. There

are more than 20 such syndromes, identified by

specific clinical characteristics

and the setting in which

they occur, particularly the provoking factors.

Simple blue breath holding attacks, usually after

the age of 6 months, provoked by frustration or other

upset, consist of a positive Valsava manoeuvre,

cyanosis, stiffening, and coma. White reflex asystolic

attacks may start earlier, usually follow minor injury,

and consist of vagal asystole, pallor, rapid coma, stiffen-

ing, and opisthotonos. Faints while standing or

sometimes from sitting are usually well recognised. All

three conditions may be accompanied by late true epi-

leptic phenomena of jerking secondary to cerebral

hypoxia or ischaemia, but their natural history remains

self limiting and harmless like that of the first episode.

Such episodes are, however, frightening for carers.

Non-epileptic attacks may be misdiagnosed and

treated with antiepileptic drugs, the attacks tending to

persist despite treatment and good prognostic features

(see table 1).

Management of epilepsy

The principles of management of epilepsy fall easily

into a problem oriented approach:

x Take detailed description of seizures

x Draw up a problem list and include some

assessment of cognition

x Clarify the problem list by doing investigations

x Decide on the need for treatment

x Predict an outcome

x Review predicted outcome at a suitable time and the

problem list annually, including a description of a typi-

cal day in the child’s life.

A description of the seizures is the main diagnostic

evidence, and a systemised approach to collecting

clinical data is required (box on next page). Several

classifications of epileptic seizures, including that of the

International League Against Epilepsy, exist but

Summary points

Identify non-epileptic attacks

Identify benign and malignant syndromes

Use predicted prognosis to guide management

Identify and manage coexisting disabilities

Ensure multidisciplinary and multiagency

coordination of treatment

Select cases amenable to surgery as soon as

possible

Neurosciences Unit,

Wolfson Centre,

Institute of Child

Health and Great

Ormond Street

Hospital for

Children NHS

Trust, University

College London

Medical School,

London WC1N

2AP, and St Piers,

Lingfield, Surrey

BGRNeville,

professor of paediatric

neurology

BMJ 1997;315:924–30

924 BMJ VOLUME 315 11 OCTOBER 1997

contain hybrid data on both seizure types and epilepsy

syndromes. The box shows a classification of seizures in

which judgment about the degree of disturbance of

consciousness is left to a second stage without initial

separation into simple and complex partial seizures.

The main advance in understanding childhood

epilepsies and their logical investigation and treatment

has been the delineation of epilepsy syndromes and

their separation into benign and malignant prognoses.

In my examples, however, I emphasise that benign may

be relative. Many children’s seizure disorders do not fit

a described syndrome, but the prognostic factors in

table 1 may be used to predict an outcome.

The genetic localisation of seizure syndromes is

separable into pure lesionless conditions that have epi-

lepsy as their only initial manifestation and conditions

with lesions that commonly cause epilepsy but which

have additional primary manifestations. The primary

pure genetic epilepsies with simple mendelian inherit-

ance and gene localisation

are benign familial neona-

tal convulsions, benign familial infantile convulsions,

partial epilepsy with auditory symptoms, autosomal

dominant nocturnal frontal lobe epilepsy, and progres-

sive epilepsy with mental retardation (Northern

epilepsy). Possible localisations are proposed for three

epilepsies with complex inheritance: febrile convul-

sions, juvenile myoclonic epilepsy, idiopathic general-

ised epilepsy.

Neonatal seizures

Benign and malignant epilepsy syndromes are both

seen in the neonatal period. Despite an overall poor

prognosis for both survival and future impairments

(largely cognitive and motor), separating the syn-

dromes into those with a comparatively good outcome

and those with a poor outcome is useful. Those with a

good outcome include benign familial neonatal

seizures, fifth day seizures,

and seizures secondary to

hypocalcaemia and subarachnoid haemorrhage.

Those with a poor outcome include seizures secondary

to hypoxic ischaemic encephalopathy and develop-

mental brain defects. Thus if neonatal seizures are age

limited without persisting neurological abnormality,

the prognosis is excellent and further investigation and

long term antiepileptic treatment are not required.

Two rare treatable recessively inherited disorders

—

pyridoxine dependent epilepsy and biotinidase

deficiency

—

may cause spontaneous seizures in the first

two years of life. All babies with unexplained epilepsy

should receive pyridoxine 100 mg/day orally, the

initial dose being given in hospital in case of a

hypotonic reaction,

and a trial of biotin or

measurement of biotinidase.

Benign syndromes of epilepsy

Benign syndromes are age limited, stereotyped, and

often provoked or provokable; they occur in otherwise

normal children and are commonly genetic.

Febrile convulsions

Febrile convulsions are epileptic seizures provoked by

fever of extracranial infective origin that occur in at

least 3% of children, mainly between the age of 6

months and 5 years. Intracranial infection must be

excluded, particularly meningitis, which is indicated by

drowsiness and irritability, and fever may show under-

lying epileptogenic disease. The seizures are usually

brief, bilateral clonic, or tonic-clonic attacks. About

10% of children have complex, prolonged, or focal sei-

zures requiring emergency treatment for status epilep-

ticus. Severe sequelae such as hemiplegia are now

uncommon, but there has been much dispute about

the risks of continuing epilepsy. The collaborative peri-

natal project showed that the risk of subsequent

epilepsy and non-febrile seizures was 10% even with all

Description of seizures

Setting or provocation

Warning or early phase (“aura”)

Motor aspects:

Type of movement

Distribution

Sensory aspects:

Type

Distribution

Responsiveness

Length

Colour change

Recovery phase

What was it like for child and carer?

Table 1 Good and adverse prognostic features for outcome of epilepsy

Good outcome Adverse outcome

Single seizure type Multiple seizure types

No additional impairment Additional neurological impairment (especially in

cognitive function)

Late age of onset Early age of onset (for the syndrome)

Episode is related to illness with full recovery or was

provoked

Spontaneous seizures

Short seizures Status epilepticus

Low rate of seizures High rate of seizures

(Good response to antiepileptic drugs) (Poor response to antiepileptic drugs)

Classification of seizures

Partial with early motor phenomena (described)

—

for

example, tonic-clonic seizures and automatic

movements

Partial with early aphasic phenomena (described)

Partial with early sensory phenomena

(described)

—

visual, limb tingling, auditory, gustatory

Partial with early psychological phenomena

(described)

—

for example, fear

Partial with loss of awareness

Partial with postictal phenomena (described)

Partial with secondary generalisation (described)

Generalised tonic-clonic:

With focal features

Without focal features

Generalised tonic

Infantile spasms:

With focal features

Without focal features

Myoclonic:

Focal (described)

Axial

Atonic

Absence seizures

Simple

With additional or atypical features

Clinical review

925BMJ VOLUME 315 11 OCTOBER 1997

three adverse factors (family history of epilepsy,

pre-existing neurological abnormality, and complex

initial seizure).

Surgical

and magnetic resonance

imaging studies,

however, both showed a high rate of

previous febrile status epilepticus in mesial temporal

sclerosis, and a causative pathogenic sequence is

supported by animal studies.

The low individual risk

from epidemiological studies

could nevertheless, in

this common disorder, produce 100-200 cases of

mesial temporal sclerosis a year in the United

Kingdom. There have been several reviews of these

data.

The practical conclusion was that febrile status

should if possible be prevented and vigorously treated.

Febrile status epilepticus is, however, commonly

the first seizure,

and subsequent preventive measures

may have little effect. Both phenobarbitone and

sodium valproate reduce the rate and severity of febrile

seizures; sodium valproate is preferred because pheno-

barbitone has behavioural side effects. The choice is

between continuous prophylaxis and the contingent

use

—

that is, at onset of seizures

—

of rectal diazepam for

those with a history of complex or frequent febrile

convulsions. Rectal diazepam is the commonest

treatment. Prophylactic management of fever may be

effective. Attending to why a child has many febrile

illnesses is also important.

Benign epilepsy with Rolandic foci

Benign epilepsy with Rolandic (centrotemporal

interictal) foci has an onset between 2 and 12 years, is

maximal between 7 and 10 years, and stops by 13 years.

The attacks are clonic, partially sensorimotor affecting

the face, bulbar (tongue and pharyngeal) muscles,

hand, and arm, and they occur particularly on waking.

An electroencephalogram shows characteristic pat-

terns. The children are free of major impairments but

may have neuropsychiatric abnormalities.

Car-

bamazepine is commonly used if seizures are unpleas-

ant. When adverse features are present (table 1),

imaging is required to exclude a Rolandic lesion.

Benign occipital epilepsy

Benign occipital epilepsy with sharp and slow wave

activity suppressed by eye opening is problematic.

Some children with typical partial seizures with visual

phenomena, postictal headache, and migrainous

characteristics

have a benign prognosis, but others

have a poorer cognitive and seizure outcome, and

some have lesions. The rules on the importance of

cognitive function in predicting the outcome of

seizures apply (table 1).

Early myoclonic epilepsies

Separation of early myoclonic epilepsies into benign

and malignant outcomes may not be possible early.

Absence seizures

Typical absence seizures are uncommon. Their separa-

tion from daydreaming is clinical. Absence seizures

totally interrupt activities in which the child is engaged

for several seconds. The outcome, mainly the rate of

tonic-clonic seizures in adult life, can be predicted by

the expected adverse factors: history of tonic-clonic

seizures, an intelligence quotient (IQ) of < 90, and a

family history of seizures. Seizures continue in about

10% of cases with no adverse factors or only one, and

in all cases with three adverse factors. Atypical and

complex absence seizures may have in addition clonic,

tonic, or automatic movements and an altered

sensorium before or after. Electroencephalograms

help diagnosis and treatment. Phenytoin and car-

bamazepine are unhelpful or aggravate the condition,

particularly myoclonic absence seizures. Absence

seizures range from a benign self limiting condition to

part of a malignant syndrome of multiple seizure types,

particularly the Lennox-Gastaut syndrome.

Malignant syndromes of epilepsy

Although developmental arrest and regression as a

consequence of epilepsy are common in malignant

syndromes of epilepsy, investigations for metabolic,

degenerative, and structural brain disease are essential.

The most severe example is early myoclonic encepha-

lopathy, which usually presents neonatally with

fragmentary myoclonus and other seizure types, with

later massive myoclonic and tonic seizures. Electro-

encephalography shows suppression or burst activity,

and one cause is non-ketotic hyperglycinaemia.

Mortality is 50%, and surviving children are severely

disabled.

Infantile spasms

Infantile spasms, or West’s syndrome, consist of runs of

tonic spasms occurring every 5-10 seconds. The

commonest spasms are flexion, but extensor and

mixed types are seen and may show lateralisation. Par-

ents may notice loss of visual and social interaction

rather than spasms. West’s syndrome is symptomatic of

many static cortical diseases, and with modern imaging

and a careful history, only 10% remain without

antecedent problems.

Attacks start in the first year

with a peak onset at around 4 months, the outcome

being worse with early onset. A grossly disorganised

epileptic electroencephalogram (hypsarrhythmia) is

characteristic but not always seen.

The mainstays of treatment have been cortico-

steroids (tetracosactrin or prednisolone), but recent

studies have shown vigabatrin to be effective,

although a comparative study has yet to be performed.

West’s syndrome is sometimes a secondary generalised

epilepsy with obvious lateralised lesions; some studies

using positron emission tomography have shown cold

(hypometabolic) lesions, which have occasionally been

excised with relief of seizures.

The occurrence of the

syndrome is no bar to the surgical treatment of focal

epilepsy, but the developmental outcome will be

adversely affected if the syndrome is present

—

for

example, the generally poor cognitive outcome and

autistic features of children with tuberous sclerosis and

other diseases with early onset seizures, including

West’s syndrome.

A death rate of about 20% and the

occurrence of cerebral palsy in 30-50% and cognitive

disability in up to 85% should prompt early ascertain-

ment, energetic treatment, and further research.

Chevrie and Aicardi showed the poor developmen-

tal outcome for most children with spontaneous

seizures

—

that is, seizures that are not provoked by

fever, infection, trauma, or metabolic disturbance

—

they start in the first year of life, with or without West’s

syndrome.

Thus all children with early onset sponta-

neous seizures should be urgently referred. The

Clinical review

926 BMJ VOLUME 315 11 OCTOBER 1997

non-spontaneous group includes non-accidental head

injury, meningitis, and hypernatraemic dehydration, so

only simple febrile convulsions may be managed in

primary care.

Severe myoclonic epilepsy of childhood

Severe myoclonic or polymorphic epilepsy begins at

4-10 months, often with a prolonged partial clonic sei-

zure associated with moderate fever. Later attacks

include multiple seizure types, including complex

absences, myoclonias, complex partial seizures, and

apnoeic attacks. With the onset of multiple types of sei-

zure, developmental regression, often severe, occurs.

Lennox-Gastaut syndrome

In the Lennox-Gastaut syndrome, multiple seizures

occur in the first eight years of life. A single seizure type

including infantile spasms may occur beforehand. The

seizures include tonic attacks (particularly in sleep),

atypical absence seizures, atonic and myoclonic attacks,

and episodes of non-convulsive status, with long

periods of impairment of motor and cognitive

function. Half of the children affected have primary

developmental delay, and all have learning problems

after five or more years of the condition; only up to

10% have a reasonable outcome. Characteristic

electroencephalograms show bilateral variable slow

(1-2 Hz) spike-waves.

Seizures are often intractable to antiepileptic drugs,

but benzodiazepines, sodium valproate, lamotrigine,

corticosteroids, and a ketogenic diet are regularly used.

Drug intoxication may contribute to disability. Fel-

bamate has been used with good effect, but its use is

restricted because of side effects. The underlying

pathology is either unknown or closely overlaps that of

infantile spasms. Occasionally resective surgical treat-

ment can be offered, and the palliative procedure of

callostomy may reduce severe drop attacks (axial myo-

clonic or atonic seizures that cause sudden falls and

head injury). The Lennox-Gastaut syndrome illustrates

the apparently insurmountable problems of providing

a coherent medical, social, and education programme

for the children and families who face dealing with

severe epilepsy and continuously varying levels of

function. Many of these children do not attend educa-

tion unless special multiagency help is available.

Partial seizures of lesional origin

Partial seizures of lesional origin are diverse, and surgi-

cal treatment may be offered. The definitive study of

the natural history of partial seizures of temporal lobe

origin predates modern imaging.

21 22

A mortality of at

least 0.5% per annum and an outcome of a third with

full recovery, a third with continuing problems, and a

third being totally dependent was reported. The

predictors of poor outcome were similar to those in

table 1 and are apparent early, giving strong

justification for surgical intervention in childhood

rather than later in adult life, as is still often the case.

Younger children, however, present with a wide range

of clinical syndromes, including early cognitive and

autistic regression with developmental tumours of the

temporal lobe, which may constitute a surgical

emergency. Clinically significant recovery from autistic

features may be achieved.

Landau-Kleffner syndrome

The Landau-Kleffner syndrome is uncommon. Typi-

cally, partial or generalised seizures (including atypical

absence seizures) begin and language comprehension

and speech are lost after two or more years of normal

development. Aphasia is not related to frequency or

severity of seizures, and clinical seizures may be absent

in about 20% of children, but they will show typical

language regression and electroencephalographic

changes (variable temporal or temporoparietal multi-

focal spikes or spike-wave complexes intensified in

sleep, including non-convulsive status epilepticus).

The range of impairments when early aphasia,

epilepsy, and typical electroencephalographic abnor-

malities are present includes behavioural disorders and

global cognitive, motor,

and social regression with

many autistic features. The contribution of this epilep-

tic mechanism to autism is not resolved. Structural

imaging usually gives normal results, and lesional vari-

ants are mentioned earlier. Antiepileptic drugs often

have no effect on the cognitive deficit, but high doses of

sodium valproate, often in combination with lamot-

rigine, sometimes help. A clinically significant, some-

times dramatic, response to corticosteroids makes

them essential, but side effects and long term depend-

ency may be a problem. Preliminary results of surgical

treatment using multiple subpial transections suggest

that about half of those with clear evidence of a driving

hemisphere can be helped.

The Landau-Kleffner syndrome is a good model

for cognitive arrest and deterioration in epilepsy. One

aspect of its pathophysiology is that part of the tempo-

ral lobe is chronically hypoperfused and hypometa-

bolic.

Thus two pathogenic mechanisms seem to

exist, one fast and electrical,producing clinical seizures,

and the other with a slower and vascular or metabolic

component, related to non-convulsive status in sleep

and producing prolonged cognitive deficits.

General features

Several of these conditions (infantile spasms, polymor-

phic epilepsy, the Lennox-Gastaut syndrome, and the

Landau-Kleffner syndrome) illustrate the degenerative

effects of severe epilepsy with loss of cognitive and

social functions but without change in magnetic

resonance images or rate of head growth. The general

designation of an epileptic encephalopathy seems jus-

tified and prompts further research into these

potentially remediable but currently severely disabling

conditions. The question posed by Sadzot’s recent edi-

torial, “Do seizures beget seizures?”

is perhaps not the

main issue in childhood when cognitive arrest and

dementia are higher research priorities.

The problems faced by families dealing with these

syndromes of epilepsy has promoted the development

of several excellent parent support groups.

Drug treatment

There are few well controlled studies and no placebo

controlled studies to justify much of how childhood

epilepsy is managed with antiepileptic drugs. Thus

arguments about whether to treat after the first, second,

or third seizure are based on a general approach rather

than evidence. The case for not necessarily treating

mild seizures and for inpatient withdrawal of all drugs

Clinical review

927BMJ VOLUME 315 11 OCTOBER 1997

in poorly controlled epilepsy has been strongly argued.

The suggested approach is to use the predicted natural

history as the main guide to treatment. A randomised,

controlled study of single drug treatment in children

aged 3 years and over with partial and generalised

tonic-clonic seizures showed similar rates of control

and side effects with phenytoin, carbamazepine, and

sodium valproate.

The study used slowly increasing

increments of drugs and blood concentrations up to

the top half of a defined therapeutic range. Phenobar-

bitone was removed early because over half the

children had unacceptable behavioural side effects.

Because of the predictive importance of previous

neurological, particularly cognitive impairments, chil-

dren were stratified for neurodisability, and this proved

to be a significant predictor of seizure control and a

crucial element of such paediatric studies.

Table 2 shows the drugs used in childhood epilepsy.

Drug licensing does not cover all of the current uses in

children, and parents should be counselled appropri-

ately. My approach to treatment is:

x To decide on the need for treatment on the basis of

prognosis

x To select a single drug and increase the dosage

slowly, giving the family clear information on how to

manage status epilepticus

x To maximise the drug dosage if seizures continue

x To add a second drug if epilepsy is not controlled,

with combined treatment lasting for sufficient time to

assess the second drug as a possible line of treatment

before withdrawal of the first

x To try further drugs and instigate detailed investiga-

tions if the epilepsy proves intractable to two drugs. If

adverse factors are absent the diagnosis of epilepsy

should be questioned

x Not to react to a brief cluster of seizures if the

epilepsy is generally well controlled

x To investigate urgently deterioration in any domain

of function.

The use of new antiepileptic drugs in children has

been recently reviewed by Pellock.

Ketogenic diets

have been in use in epilepsy for 75 years and come in

two forms: using animal fat and, more recently,

medium chain triglycerides. Several reports support

their use,

but well controlled studies of comparative

(with antiepileptic drugs) and selective (which type of

epilepsy) efficacy are not available.

Although the concomitant use of psychotropic

drugs with antiepileptic drugs in epilepsy needs careful

monitoring, behavioural and drug treatment of hyper-

activity and attention deficit can be successful, but this

issue requires further study.

Surgery

The surgical treatment of childhood epilepsy is

growing rapidly, justified by studies of the natural

history of the disease, which show that waiting for

remission in most children with intractable epilepsy is

a false hope. If a surgical solution is possible it should

Table 2 Drugs of use in different types of childhood epilepsy

Dose

(increments,

when applicable)

Usual

maximum

daily dose

(mg/kg/day)

Blood

reference range

Generalised tonic-clonic and focal seizures

Carbamazepine 4 (4) mg/kg/day 20 Up to 50 ìmol/l (epoxide

(metabolite) up to 9 ìmol/l)

Sodium valproate 10 (5) mg/kg/day 40 Up to 700 ìmol/l

Phenytoin 5 (2) mg/kg/day 10 Up to 80 ìmol/l

Vigabatrin 30 (10) mg/kg/day 80 40-270 ìmol/l

Phenobarbitone 3 (2) mg/kg/day 5 Up to 170 ìmol/l

Primidone 10 (5) mg/kg/day 20 Up to 60 ìmol/l

Lamotrigine:

Alone 1 (2) mg/kg/day 10 4-16 ìmol/l*

With sodium valproate 0.1 (0.1) mg/kg/day 1.5 4-16 ìmol/l*

Gabapentin† 5 mg/kg/day (5 daily) 70 Up to 120 ìmol/l

Topiramate† 0.5-1 (1) mg/kg/day 4-5 6-74 ìmol/l

Prednisolone‡ 2 mg/kg/day NA NA

Ketogenic diet‡ NA NA NA

Chlormethiazole 10-20 mg/kg/day NA 3-18 mg/l

Absence seizures

Ethosuximide 10 (5) mg/kg/day 30 700 ìmol/l

Sodium valproate 10 (5) mg/kg/day 40 Up to 700 ìmol/l

Lamotrigine 1 (2) mg/kg/day 10 4-16 ìmol/l*

Myoclonic seizures§

Nitrazepam 0.1-0.5 mg/kg/day 0.5 50-150 ìg/l

Clonazepam 0.05-0.2 mg/kg/day 0.2 25-85 ìg/l

Clobazam 0.25-1.0 mg/kg/day 1.0 Up to 200 ìg/l (desmethyl

clobazam (metabolite) up to

2000 ìmol/l)

Sodium valproate 10 (5) mg/kg/day 40 Up to 700 ìmol/l

Lamotrigine 1 (2) mg/kg/day 10 4-16 ìmol/l*

Febrile seizures

Intermittent diazepam

Sodium valproate 10 (5) mg/kg/day 40 Up to 700 ìmol/l

Phenobarbitone 3 (2) mg/kg/day 5 Up to 170 ìmol/l

Antibiotics

Status epilepticus

Diazepam Intravenous or rectal

0.15-0.3 mg/kg

NA 1 mg/l (desmethyl diazepam

(metabolite) up to 1.5 mg/l)

Lorazepam Intravenous 0.03-0.05 mg/kg NA 1.5 mg/l

Phenytoin Intravenous 15-20 mg/kg (slowly) NA Up to 80 ìmol/l

Chlormethiazole Up to 0.7 g/h (adult dose),

300-800 mg in 100 ml (4 ml/min)

(5-18 mg/kg/h)

NA 3-18 mg/l

Thiopentone 4-8 mg/kg (induction),

10 mg/kg/h (ventilation)¶

NA NA

Infantile spasms

Tetracosactrin depot 0.5 mg 1 mg NA

Prednisolone 2 mg/kg/day for 2 weeks NA NA

Vigabatrin Up to 150 mg/kg/day 150 40-270 ìmol/l

Nitrazepam 0.1-0.5 mg/kg/day 0.5 50-150 ìg/l

Sodium valproate 10 (5) mg/kg/day 40 Up to 170 ìmol/l

NA=not applicable. *To be revised upwards. †Not licensed for use in children. ‡Also for intractable

myoclonic epilepsy and absence seizures. §Do not use carbamazepine, phenytoin, or vigabatrin.

¶Not licensed for use in children under 2 years.

Indications for and types of surgery

Indications for assessment

Intractable partial seizures

Intractable hemiepilepsy

Intractable, apparently generalised, epilepsy with focal

clinical electroencephalographic or radiological

features

Regression in epilepsy

Procedures

Lesionectomy

Focal resections:

Temporal

Extratemporal

Hemispherectomy

Functional procedures:

Callostomy

Subpial transections

Clinical review

928 BMJ VOLUME 315 11 OCTOBER 1997

therefore be used early in the child’s decline. The main

advances allowing these developments have been

magnetic resonance imaging,

T2 mapping,

proton

spectroscopy,

single photon emission tomography

between and during seizures,

and video telemetry.

The use of these magnetic resonance techniques in

children requires sedation, anaesthesia, and age appro-

priate sequences and normative values. The complex-

ity of these children’s problems requires a team of

paediatric neurologists, neurosurgeons, anaesthetists,

neuropsychologists, psychologists,psychiatrists, speech

therapists, nurses, and neuropathologists.

The use of surgery for epilepsy is developing

rapidly.

Part of paediatric practice is the young end of

adult series, which is dominated by temporal lobe

resections for mesial temporal sclerosis (fig 1). In one

study of surgery in children, half of the subjects were 6

and younger, a group that usually has behavioural and

cognitive impairments.

Most would not enter adult

surgical studies because mortality and morbidity are

high in children. The younger children’s problems

include:

x Unilateral congenital hemispheric defects (particu-

larly developmental migration defects) (fig 2) and

ischaemic intrauterine events with hemiplegia and a

preserved hemisphere with early onset epilepsies,

including infantile spasms. Affected children may have

total arrest of development, which may be related as

much to subclinical seizures as obvious epilepsy

x Acquired damage due to injury and infection, which

suggests the possibility of damage more widespread

than the source of the seizure

x The Sturge-Weber syndrome, in which an acute,

combined epileptic and vascular encephalopathy leads

to acquired unilateral damage but global cognitive

regression (fig 3)

x Focal areas of dysplasia, including dysembryoplastic

neuroepithelioma, which in the temporal lobe may be

associated with massive regression and autistic features

(fig 4)

x Acquired unilateral inflammatory disease (Rasmus-

sen’s syndrome)

x Acquired aphasia (Landau-Kleffner syndrome).

The clinical indications for surgery and the types of

procedures are given in the box. A non-invasive pack-

age of investigations has allowed most children to be

evaluated without invasive monitoring. Invasive moni-

toring is required, however, when surgery might

encroach on eloquent cortex or the extent and site of

epileptogenesis is unclear.

The wide range of disease, epilepsy syndromes, and

additional disabilities in childhood epilepsy make it

Fig 1 Right hippocampal sclerosis Fig 2 Left hemimegalencephaly

Fig 3 Gadolinium enhanced magnetic resonance scan showing left

cerebral atrophy and pial enhancement pathognomic of Sturge-Weber

syndrome

Fig 4 Right cortical dysplasia

Clinical review

929BMJ VOLUME 315 11 OCTOBER 1997

essential that a scheme of categorisation of predicted

outcome is used. Such an approach, proposed by Tay-

lor et al, uses three categories: expected cure, predicted

clinically significant amelioration, and procedures of

less certain outcome.

Taylor et al also develop a con-

tract with the parents and patient when appropriate,

exploring their aims for surgery and deciding whether

these are achievable and with what risks. This is of

course a basic requirement of all medical interventions.

Several relevant parent support groups are part of Contact-a-

Family, 170 Tottenham Court Road, London W1P 0HA (tel:

0171 383 3555).

I thank Dr P Patsalos of the Pharmacology and Therapeutics

Unit of the Institute of Neurology and National Hospital for

Neurology and Neurosurgery, London, for providing the blood

reference ranges given in table 2. I also thank Professor J

Aicardi, Dr J H Cross, and Dr R Scott for helpful suggestions.

1 Aicardi J. Epilepsy in children. 2nd ed. New York: Raven, 1994.

2 Stephenson JBP. Fits and faints. London: MacKeith Press with Blackwell

Scientific, 1990. (Clinics in developmental medicine No 9.)

3 Elmslie F, Gardiner M. The epilepsies. In: Rimoin DL, Connor JM, Pyeritz

RE, eds. Emery and Rimoin’s principles and practice of medical genetics.New

York: Churchill Livingstone, 1996:2177-96.

4 Pryor DS, Don N, Macourt DC. Fifth day fits: a syndrome of neonatal

convulsion. Arch Dis Child 1981;56:753-8.

5 Kroll JS. Pyridoxine for neonatal seizures: an unexpected danger. Dev

Med Child Neurol 1985;27:377-9.

6 Nelson KB, Ellenberg JH, eds. Febrile seizures. New York: Raven, 1981.

7 Nelson KB, Ellenberg JH. Prenatal and perinatal antecedents of febrile

seizures. Ann Neurol 1990;27:127-31.

8 Falconer MA, Taylor DC. Surgical treatment of drug resistant epilepsy

due to mesial temporal sclerosis. Arch Neurol 1968;19:353.

9 Cross JH, Jackson GD, Neville BGR, Connelly A, Kirkham FJ, Boyd SG, et

al. Magnetic resonance imaging in intractable epilepsy of childhood. Arch

Dis Child 1993;69:104-9.

10 Meldrum BS, Horton RW, Brierley JB. Epileptic brain damage in adoles-

cent baboons following seizures induced by allylglycins. Brain

1974;97:407.

11 Verity C, Ross E, Golding J. Outcome of childhood status epilepticus and

lengthy febrile convulsions: findings of a national cohort study. BMJ

1993;307:225-8.

12 Liu Z, Mikati M, Holmes G. Mesial temporal sclerosis: pathogenesis and

significance. Paediatric Neurol 1995;12:5-16.

13 Beaussart M. Benign epilepsy of children with Rolandic (centro-

temporal) paroxysmal foci: a clinical entity. Study of 221 cases. Epilepsia

1972;13:795-811.

14 Gastaut H. Benign epilepsy of childhood with occipital paroxysms. In:

Roger J, Bureau M, Dravet C, Dreifuss FE, Perret A, Wolf P, eds. Epileptic

syndromes in infancy, childhood and adolescence. 2nd ed. London: Libbey,

1992:201-17.

15 Hrachovy RA, Frost JD. Infantile spasms. Pediatr Clin North Am

1989;36:311-30.

16 Chiron C, Dulac O, Beaumont D, Palacios L, Pajot N, Mumford J. Thera-

peutic trial of vigabatrin in refractory infantile spasms. J Child Neurol

1991;6:2S52-6.

17 Chugani HT, Shields WD, Shewmon DA, Olson DM, Phelps ME, Peacock

WJ. Infantile spasms. I. PET identifies focal cortical dysgenesis in

cryptogenic cases for surgical treatment. Ann Neurol 1990;27:406-13.

18 Gomez MR, ed.Tuberous sclerosis. 2nd ed. New York: Raven, 1988.

19 Chevrie JJ, Aicardi J. Convulsive disorders in the first year of life:

neurological and mental outcome and mortality. Epilepsia 1978;19:67-74.

20 Cavazzuti GB, Ferrari P, Lalla P. Follow-up of 482 cases with convulsive

disorders in the first year of life. Dev Med Child Neurol 1984;26:425-37.

21 Ounsted C, Lindsay J, Richards P. Temporal lobe epilepsy 1948-1986: a bio-

graphical study. Oxford: MacKeith Press with Blackwell Scientific,1987.

22 Taylor DC. Mental state and temporal lobe epilepsy: a correlative account

of 100 patients treated surgically. Epilepsia 1972;13:727-76.

23 Neville BGR, Harkness WFJ, Cross JH, Cass HC, Burch VC, Lees J, et al.

Surgical treatment of autistic regression. Paediatric Neurol 1997;16:

137-40.

24 Beaumanoir A. The Landau-Kleffner syndrome. In: Roger J, Dravet C,

Bureau M, Dreifuss FE, Perret A, Wolf P, eds. Epileptic syndromes in infancy,

childhood and adolescence. London: Libbey, 1993:231-43.

25 Neville BGR,Boyd SG. Selective epileptic gait disorder. J Neurol Neurosurg

Psychiatry 1995;58:371-3.

26 Morrell F, Whisler WW, Smith MC, Hoeppner TJ, de-Toledo-Morrell L,

Pierre-Lewis SJ, et al. Landau-Kleffner syndrome. treatment with subpial

intracortical transection. Brain 1995;118:1529-46.

27 Maquet P, Hirsch E, Metz-Lutz MN, Motte J, Dive D, Marescaux C, et al.

Regional cerebral glucose metabolism in children with deterioration of

one or more cognitive functions and continuous spike-and-wave

discharges during sleep. Brain 1995;118:1497-520.

28 Sadzot B. Epilepsy: a progressive disease? BMJ 1997;314:391.

29 de Silva M, MacArdle B, McGowan M, Hughes E, Stewart J, Neville BGR,

et al. Randomised comparative monotherapy trial of phenobarbitone,

phenytoin, carbamazepine, or sodium valproate for newly diagnosed

childhood epilepsy. Lancet 1996;347:709-13.

30 Pellock JM. Utilization of new antiepileptic drugs in children. Epilepsia

1996;37(suppl I): S66-73.

31 Prasad AN, Stafstrom CF, Holmes GL. Alternative epilepsy therapies: the

ketogenic diet, immunoglobulins and steroids. Epilepsia 1996;37(suppl

I):S81-95.

32 Jackson GD, Connelly A, Duncan JS, Gadian GD, Grunewald R. MRI

detection of hippocampal pathology in temporal lobe epilepsy: increased

sensitivity using quantitative T

relaxometry. Neurology 1993;43:1973-99.

33 Gadian DG, Connelly A, Duncan JS, Cross JH, Kirkham FJ, Johnson CL,

et al.

H magnetic resonance spectroscopy in the investigation of “intrac-

table epilepsy.” Acta Neurol Scand 1994;89 (suppl):116-22.

34 Cross JH, Gordon I, Jackson GD, Boyd SG, Todd-Pokrapek A, Anderson

PJ, et al. Children with intractable focal epilepsy: ictal and interictal

99m

HMPAO single photon emission computed tomography. Dev Med Child

Neurol 1995;37:673-81.

35 Engel J Jr, ed. Surgical treatment of the epilepsies. 2nd ed. London: Raven,

1993.

36 Duchowny M. Epilepsy surgery in children. Curr Opin Neurol

1995;8:112-6.

37 Taylor DC, Cross JH, Neville BGR. Mental states of childhood epilepsy

surgery candidates. Epilepsia 1996;37(suppl 4):51.

When I use a word .. .

Fingerprints

Fond though doctors are of using long words of Greek and Latin

origins, in some cases we use ordinary English words. We say

hiccup (not singultus), yawn (not pandiculation), and fingerprints

(not dermatoglyphics).

The recent discovery that koala bears have fingerprints that more

closely resemble those of humans than chimpanzees’ fingerprints

do is said to support the hypothesis that fingerprints evolved as an

aid to climbing. Whether or not that is so, fingerprints have been

used for centuries as means of identification. Greek vases, for

instance, bear examples as signatures, and in India a fingerprint

used in this way by an illiterate person was known as tipsahi. But

fingerprints were first systematically described by Johannes

Evangelista Purkenje (sic) in his thesis, Commentatis de examine

physiologico organi visus et systematis cutanei, published in Breslau in

1823. And they were really put on the map by the astronomer

William Herschel (the Younger), who first devised a method for

printing them in 1858, and by Francis Galton the geneticist, who in

1892 wrote a book, Fingerprints, about the differences in skin

creases in different individuals.

The word dermatoglyphics was invented in 1926 by Harold

Cummins and Charles Midlo and was used for the first time in a

paper on what they called “epidermal ridge configurations” (Am J

Phys Anthropol 1926;9:471-502), where they restricted its use to

ridges and their arrangements, excluding flexion creases and

other secondary folds. Its origin is simple: from ä´åñìá (derma),

the skin, and ãëu¨õöù (glupho), I sculpt.

In everyday speech we may prefer to say “fingerprints,” but

“dermatoglyphics” is used much more often in publications,

partly no doubt because it sounds more scientific, but also

because it can be used to describe not only the prints themselves

but also the study of them. Thus, although “fingerprint” and its

derivatives were used in 3825 titles or abstracts of bioscience

papers published between 1966 and 1997, in only 197 of those

was dermatoglyphics meant, the other cases all being to do with

fingerprinting as a molecular biological tool. In contrast, 1475

papers about proper fingerprints used the word dermatoglyphics.

Dermatoglyphics also leaves a distinctive fingerprint in the

dictionary: it is the longest word in the English language (15

letters) that uses no letter twice.

Jeff Aronson, clinical pharmacologist, Oxford

Clinical review

930 BMJ VOLUME 315 11 OCTOBER 1997