Seizure 1997; 6: 159-174

Established antiepileptic drugs

MARTIN J. BRODIE & MARC A. DICHTER

Epilepsy Unit, University Department of Medicine and Therapeutics, Western Infirmary, Glasgow,
Scotland, UK, and Department of Neurology, Graduate Hospital and University of Pennsylvania,
Philadelphia, USA

Correspondence to: Professor M.J. Brodie, Epilepsy Unit, Department of Medicine and Therapeutics,
Western Infirmary, Glasgow Gil 6NT, Scotland, UK

Despite the recent entry into the market-place of a range of new pharmacologicaltreatments for epilepsy, most
patients still receive the standard antiepileptic drugs.This review considersthe clinical place and practical useof
these agents.Detailed consideration is given to carbamazepine,phenytoin, sodium valproate, phenobarbital and
ethosuximide, with lesser emphasison primidone, clobazam and clonazepam. Individualization of therapy,
polypharmacy, refractory epilepsy, therapeutic drug monitoring, pregnancy, withdrawing treatment, epilepsy
prophylaxis and referral to an epilepsy centre are also discussed. The paper concludes with a statement of 12 basic
rules in prescribingestablishedantiepileptic drugs.

Key words: antiepileptic drugs; epilepsy; interactions; pharmacokinetics:side-effects.

INTRODUCTION tic drug therapy is usually begun when the patient

Around 50 million people have epilepsy. The
annual incidence varies between 20 and 70
individuals per 100,000’ with a point prevalence
of o.4-o.8%2. Incidence rates are highest in
childhood, plateau from 15 to 65 years, and rise
again in the elderly. Overall, 5% of the world’s
population will have a seizure at some time in
their lives. These figures exclude febrile convul-
sions, which manifest in approximately 5% of
children. They do, however, support a substantial
remission rate that can occur in patients who have
never received treatment with an antiepileptic
drug3. There is evidence, too, that the more
seizures a patient experiences before treatment is
begun, the more likely the epilepsy is to prove
refractory4. This is likely to be a consequence of
the severity of the underlying process5.
Seizures represent the clinical manifestation of
a number of diverse neurological and systemic
disorders in addition to a range of idiopathic (or
genetic) epilepsies. A specific cause can currently
be identified in only approximately 30% of
patients6. The diagnosis is made from the
description of the episodes and the clinical
context in which they occur, supplemented on
occasion by electronecephalography. Antiepilep-
has suffered more than one unprovoked seizure
within a year. Although the majority of treated
patients will remain fully controlled, as many a
30% will continue to suffer seizures in the face of
optimal deployment of antiepileptic medication’.
Despite the entry into the market-place of a range
of new pharmacological treatments for epilepsy,
most patients still receive the standard anti-
epileptic drugs. This review considers the current
place and practical use of these agents.
WHEN TO TREAT
Most patients with recurrent epileptic seizures
will require treatment. Seizures can cause injury
secondary to falls, excessive muscular contrac-
tions associated with convulsive activity, and
inappropriate automatisms. Their presence will
prevent an individual from driving, and may
interfere with a variety of important educational,
vocational and social activities. In addition,
although still somewhat controversial, there is
growing concern that some types of seizures may
cause brain injury, producing cognitive and
memory problems, and promoting a propensity to
further seizures. Finally, there is a low but
significant mortality associated with uncontrolled

1059-1311/97/030159+ 16 $12.00/O Q 1997 British Epilepsy Association

 160
seizures that has been estimated at approximately
O.l-0.3% per year’. Patients may die during a
seizure or may be found dead unexpectedly in the
absence of direct evidence of a seizureg.
More controversial is the issue whether to treat
a single seizure. This may be a consequence of
severe sleep deprivation, vasovagal syncope, or
drug or alcohol withdrawal. Patients suffering
such a provoked seizure do not usually have
epilepsy and do not need antiepileptic drug
treatment. Patients who experience an
unprovoked seizure, however, have a significant
chance of developing recurrent events (ranging
from 31 to 71% depending on other risk
factorsloY”) and may already have epilepsy.
Those with an underlying neurological abnor-
mality or lesion, or who have a recognized
syndrome such as juvenile myoclonic epilepsy,
are at higher risk and should be treated. There is,
in addition, growing evidence that treating a
single unprovoked tonic-clonic seizure will pre-
vent others”, although it is not likely to alter the
overall progno&.
The decision whether or not to treat a patient
with an antiepileptic drug should be made after
ample discussions of the risks and benefits of both
courses of action. The best drug for the patient’s
seizure type(s) and lifestyle should be selected
and administered in a dose that is high enough to
bring the circulating concentrations into a relative
target (therapeutic) range without producing
unacceptable side-effects. Once treatment is
instituted, the goal should be restoration of a
normal life by complete seizure control using a
single medication with no or minimal side-effects.
Common errors in treating patients include
tolerating occasional seizures in patients who
ought to be seizure-free, mistaking cognitive and
sedative side-effects of drugs for unavoidable
consequences of the epileptic disorder, and failing
to refer patients to an epilepsy clinic in a timely
fashion.
CHOICE OF DRUG
Partial seizures
Carbamazepine (CBZ), phenytoin (PHT), phen-
obarbital (PB), primidone (PRM) and valproic
acid (VPA) are all effective in reducing the
frequency of partial seizures13-14. CBZ and PHT
are better tolerated than PB, and especially
PRM13. There is some evidence that CBZ may be
more effective than VPA for this indicationr4, but
other studies have not supported this asser-
M.J. Brodie & MA. Dichter
tion1s-‘7. CBZ is regarded as the drug of first
choice for partial seizures by most epilepsy
specialists. In the USA many neurologists prefer
to start with PHT. Some clinicians, particularly in
Europe, choose to use VPA ahead of PHT
because of the latter’s propensity to saturation
kinetics and their perception of a less favourable
side-effect profile. PB is a second-line drug in
most cases, despite its being the least expensive
and most widely available, because of its ten-
dency to produce sedation and depression in
adults and hyperactivity and aggression in chil-
dren. Among the newer antiepileptic agents,
lamotrigine, gabapentin, vigabatrin, oxcar-
bazepine, felbamate, topiramate and tiagabine all
possess useful efficacy against partial seizures”.
Generalized seizures
For tonic-clonic seizures, VPA, PHT and CBZ
are all effective. VPA is regarded as the drug of
choice for patients with tonic-clonic seizures and
spike-wave discharges on the EEG and for other
forms of idiopathic generalized epilepsy, particu-
larly myoclinic jerks and absences. Ethosuximide
(ESM) still has a limited place for generalized
absences. Here the treatment should be aimed to
eliminate all three per second spike-wave dis-
charges from the EEG. For those patients in
whom absences coexist with tonic-clonic seizures
and/or myoclonic jerks, VPA should be pre-
ferred. Clonazepam (CLZ) has efficacy against
absences, myoclonic jerks and tonic-clonic sei-
zures. Its value is acutely limited by sedation and
by the rapid development of tolerance on chronic
administration. Lamotrigine appears to be
effective for all types of generalized seizures18.
INDIVIDUAL DRUGS
Substantial data have accumulated over many
decades on the clinical pharmacokinetics of the
established antiepileptic drugs (Table 1). Their
indications and a guide to dosing in children and
adults are outlined in Tables 2 and 3. These drugs
are not without their hazards and their optimum
use must be governed by an appreciation of their
potential for dose-related and idiosyncratic toxi-
city (Table 4). The clinical use of each drug will
be considered, highlighting the practical problems
Established antiepileptic drugs 161

Table 1: Pharmacokinetics of established antiepileptic drugs

Absorption Protein Elimination Route(s) of Comments
(bioavail- binding half-life elimination
ability) (% bound) (hours)

Carba- Slow absorption 70-80 24-45 (single) Hepatic metabolism Enzyme inducer
mazepine (75-85%) 8-24 (chronic) Active metabolite Metabolic autoinduction
Clobazam Rapid absorption 87-90 IO-30 Hepatic metabolism Sedative
(90-100%) Active metabolite Tolerance
Clonazepam Rapid absorption 80-90 30-40 Hepatic metabolism Sedative
(SO-90%) Tolerance
Ethosuximide Rapid absorption 0 20-60 Hepatic metabolism More rapid clearance in
(90-95%) 25% excreted unchanged children
Phenobarbital Slow absorption 48-54 72-144 Hepatic metabolism Enzyme inducer
(95-100%) 25% excreted unchanged Sedative
Tolerance
Phenytoin Slow absorption 90-93 9-40 Saturable hepatic Enzyme inducer
@S-90%) metabolism Elimination half-life
concentration-dependent
Primidone Rapid absorption 20-30 4-12 Hepatic metabolism Sedative
(90-100%) Active metabolites Phenobarbital a metabolite
40% excreted unchanged Tolerance
Sodium Rapid absorption 88-92 7-17 Hepatic metabolism Enzyme inhibitor
valproate (100%) Active metabolites Concentration-dependent
protein binding

likely to be encountered during everyday absence or myoclonic seizures2’. It acts by

administration. preventing repetitive firing of sodium-dependent
action potentials in depolarized neurones21 via
use- and voltage-dependent blockade of Na+
Carbamazepine
channels22. CBZ should be introduced at low
CBZ was synthesized by Schindler at Geigy in dosage (100-200 mg daily) to allow tolerance to
1953 in an attempt to compete with the newly develop to its central nervous system side-
introduced antipsychotic chlorpromazine. The effect?. The dose can then increased in 2-4
first clinical studies in epilepsy were not carried weekly increments to a maintenance amount that
out until 1963. Over the years CBZ has slowly controls the seizure disorder. A balance must be
gained acceptance as an important treatment for achieved between speed of seizure control and
partial and tonic-clonic seizures”. In Europe, it is acceptance of temporary central nervous system
usually preferred over PHT. CBZ is not effective, toxicity. The final dose often depends on the
and may even be deleterious, for patients with extent of autoinduction of metabolism24.

Table 2: Dosage guidelines for established antiepileptic drugs in children

Drug Indications Starting Standard Dosage
dose bg/kg/day) maintenance interval
dose (mg/kg/day)

Carbamazepine Partial and generalized tonic-clonic seizures 5 10-25 bid-qid

Clobazam Partial and generalized seizures 0.25 0.5-l od-bid
Clonazepam Myoclonic epilepsy 0.025 0.025-0.1 bid-tid
Lennox-Gastaut syndrome
Infantile spasms
Status epilepticus
Ethosuximide Generalized absences 10 15-30 od-bid
Phenobarbital Partial and generalized tonic-clonic seizures 4 4-8 od-bid
Newborn seizures
Status epilepticus
Phenytoin Partial and generalized tonic-clonic seizures 5 5-15 od-bid
Status epilepticus
Primidone Partial and generalized tonic-clonic seizures 10 20-30 od-bid
Sodium Partial seizures and generalized epilepsies 10 15-40 od-tid
valproate
 162 M.J. Brodie & M.A. Dichter

Table 3: Dosage guidelines for established antiepileptic drugs in adolescents and adults
Drugs Indications Starting Commonest Standard Dosage
dose daily dose maintenance interval
(w) bg) dose
(mg)

Carbamazepine Partial and generalized tonic-clonic 200 600 400-2000 od-qid

seizures
Clobazam Partial and generalized seizures 10 20 10-40 od-bid
Clonazepam Myoclonic and generalized tonic- 1 4 2-8 od-bid
clonic seizures
Ethosuximide Absence seizures 500 1000 500-2000 od-bid
Phenobarbital Partial and generalized tonic-clonic, 60 120 60-240 od-bid
myoclonic, clonic and tonic seizures
Status epilepticus
Phenytoin Partial and generalized tonic-clonic 200 300 100-700 od-bid
seizures
Status epilepticus
Primidone Partial and generalized tonic-clonic 250 300 250-1500 od-bid
seizures
Sodium AI1 generalized seizures 500 1000 500-3000 od-bid
valproate Partial seizures

Some patients are unable to tolerate the 120mmol/l the patient may present with con-
neurotoxic side-effects of CBZ even at low fusion, peripheral oedema and worsening seizure
dosage and concentration. Diplopia, headache, contr0119. Orofacial dyskinesias and cardiac ar-
dizziness, nausea and vomiting are the com- rhythmias are other occasional complications.
monest complaints. A contribution to these As well as inducing its own metabolism, CBZ
side-effects is likely to come from the active can accelerate the breakdown of a number of
metabolite CBZ 10,ll epoxide25. These symp- other lipid-soluble drugs29. The commonest inter-
toms provide a ceiling to dosage for many action is with the oral contraceptive pill, neces-
patients with refractory epilepsy. In addition, sitating for most women a daily oestrogen dose of
high peak plasma concentrations often result in 50 mcg or more30. Other important metabolic
intermittent side-effects around 2 hours after targets include VPA, ESM, corticosteroids, anti-
dosing, necessitating three or four times daily coagulants, antipsychotics and cyclosporin. Drugs
administration in some patients. Such problems that inhibit CBZ metabolism resulting in toxicity
can be overcome by prescribing a controlled- include PHT, cimetidine, dextropropoxyphene,
release formulation, now available widely, diltiazem, erythromycin, isoniazid, verapamil and
which can be given once26 or twice*’ daily in all viloxaxine. An important pharmacodynamic in-
patients. teraction takes place with lamotrigine, resulting
CBZ can cause a range of idiosyncratic in headache, nausea, dizziness, diplopia and
reactions, the most common of which is a ataxia which responds to a reduction in the dose
morbilliform rash in around 10% of patients”. of either drug3’. The less common neurotoxic
Slow dosage titration may reduce this figure*‘. interaction with lithium (confusion, disorienta-
Other unusual, but more severe, skin eruptions tion, drowsiness, ataxia, termor, hyperreflexia)
include erythema multiforme and Stevens- is also not associated with altered drug
Johnson syndrome. Reversible, mild leukopenia concentrations3*.
often occurs, but does not require discontinuation The substantial variation in any given in-
of therapy unless accompanied by evidence of dividual in CBZ concentrations over the course of
infection or if the white cell count slips below the day-as much as 100% with twice-daily
1 X 109/1. Blood dyscrasias and toxic hepatitis are dosing-makes the interpretation of concentra-
rarer problems. Long-term side-effects with CBZ tion monitoring problematical. In many patients,
are few. At high concentrations, the drug has an the dosage can be titrated adequately on clinical
antidiuretic, hormone-like action, that can result criteria alone33. Exceptions include those in
in fluid retention with cardiac failure in the whom compliance is suspect and patients taking a
elderly. Mild hyponatremia is usually asympto- cocktail of antiepileptic drugs like to interact
matic, but if the serum sodium falls below mutually.
Established antiepileptic drugs 163

Table 4: Side-effects of established antieoileotic drum

Carbamazepine Clobazam Clonazepam Ethosuximide

*Diplopia *Fatigue *Fatigue *Nausea

*Dizziness *Drowsiness *Sedation Anorexia
*Headache Dizziness *Drowsiness Vomiting
*Nausea Ataxia Dizziness Agitation
Drowsiness Irritability Ataxia Drowsiness
Neutropenia Aggression Irritability Headache
Hyponatraemia Hypersalivation Aggression (children) Lethargy
Hypocalcaemia Bronchorrhoea Hyperkinesia (children)
Orofacial dyskinesia Weight gain Hypersalivation
Cardiac arrhythmia Muscle weakness Bronchorrhoea
Psychosis Psychosis

*Morbilloform rash Rash Rash Rash

Agranulocytosis Thrombocytopenia Erythema multiforme
Aplastic anaemia Stevens-Johnson syndrome
Hepatotoxicity Lupus-like syndrome
Photosensitivity Agranulocytosis
Stevens-Johnson syndrome Aplastic anaemia
Lupus-like syndrome
Thrombocytopenia
Pseudolymphoma
Teratogenicity
Phenobarbital Phenytoin Primidone Sodium valproate

*Fatigue *Nystagmus *Fatigue *Tremor

*Listlessness *Ataxia *Listlessness *Weight gain
*Tiredness Anorexia *Tiredness *Hair fall
*Depression Dyspepsia *Depression Anorexia
*Insomnia (children) Nausea *Psychosis Dyspepsia
*Distractability (children) Vomiting *Decreased libido Nausea
*Hyperkinesia (children) Aggression *Impotence Vomiting
*Irritability (children) Depression *Hyperkinesia (children) Alopecia
Aggression Drowsiness *Irritability (children) Peripheral oedema
Poor memory Headache Nausea Drowsiness
Decreased libido Paradoxical seizures Vomiting Hyperammonaemia
Impotence Megoblastic anaemia Nystagmus
Folate deficiency Hyperglycaemia Ataxia
tNeonatal haemorrhage Hypocalcaemia Folate deficiency
Hypocalcaemia Osteomalacia Hypercalcaemia
Osteomalacia tNeonatal haemorrhage Osteomalacia
Megoblastic anaemia
tNeonatal haemorrhage

Macropapular rash *Acne Rash Acute pancreatitis

Exfoliation *Gum hypertrophy Agranulocytosis Hepatotoxicity
Toxic epidermal necrolysis *Coarse facies Thrombocytopenia Thrombocytopenia
Hepatotoxicity *Hirsutism Lapus-like syndrome Stupor
Frozen shoulder Blood dyscrasias Teratogenicity Encephalopathy
Teratogenicity Lupus-like syndrome Teratogenicity
Reduced serum 1 gA
Pseudolymphoma
Peripheral neuropathy
Rash
Stevens-Johnson syndrome
Dupuytren’s contracture
Hepatotoxicity
Teratogenicity
Above line: Dose-related
Below line: Idiosyncratic.
* Commonest side-effects.
t Maternal treatment.
164
Phenytoin
The appreciation that an epileptic seizure was
accompanied by an electrical ‘storm’ in the brain
led Houston Merritt and Tracy Putnam to screen
a wide of variety of phenyl compounds against
electrically induced seizures in cats in the belief
that this chemical grouping was responsible for
the anticonvulsant effect of PB. This led rapidly
to the discovery of PHT. Several mechanisms
have been implicated in explaining its antiepilep-
tic action, but voltage-and use-dependent
blockade of Na’ channels appears to be the
primary one 34. It can also reduce the amplitude of
synaptic potentials, but at low stimulus fre-
quencies this effect is relatively smal13’. PHT is an
effective treatment for partial and tonic-clonic
seizures. It is one of a handful of drugs that
switches from first-order to saturation kinetics at
therapeutic dosage36. Accordingly, at concentra-
tions around 60~mol/l (15 mg/l) a moderate
increment in dose can produce an unexpectedly
large rise in plasma level with accompanying
neurotoxicity. Conversely, the circulating con-
centration can fall precipitously when the dose is
reduced mostly, sometimes resulting in seizure
breakthrough. The dosage, therefore, producing
the same circulating concentration, varies greatly
among individuals.
A starting amount of 5 mg/kg body weight will
produce concentrations within the target range of
40-80 pmol/I (lo-20 mg/I) in most patients.
Some, however, will saturate at this dose and
present with neurotoxicity. Others will require a
much higher dose, particularly enzyme-induced
alcohol abusers. Below 32 pmol/l (8 mg/l), an
increment not exceeding 100 mg can be made;
patients with PHT concentrations above
32 pmol/l (8 pg/ml) should receive smaller do-
sage increments (no more than 50mg) with
frequent monitoring. At levels nearing the upper
limit of the target range, the dose should be
increased by as little as 25 mg daily. Some
patients will tolerate concentrations above
100 pmol/l (25 mg/l) with benefit.
PHT can produce a range of dose-related and
idiosyncratic adverse effects (Table 4). Of the
latter, cosmetic changes (gum hyperplasmia,
acne, hirsutism, facial coarsening), although often
mild, can be troublesome. Symptoms of neuro-
toxicity (drowsiness, dysarthria, tremor, ataxia,
cognitive difficulties) become increasingly likely
the more the concentration exceeds 80pmol/l
(20mg/l). Younger patients may report nausea
and vomiting. The diagnosis of PHT toxicity
should be made on clinical grounds and not be
assumed from a high concentration. The patient
M.J. Erodle & M.A. Dichter
will complain of mental slowing and unsteadiness,
and neurological examination will reveal nystag-
mus and ataxia. Heel-toe walking is particularly
sensitive. Permanent cerebellar damage3’ may be
a consequence of chronic toxicity, and so it is
important to examine the patient regularly. A
paradoxical increase in seizure frequency may
also point the way to PHT toxicity.
PHT is an enzyme inducer and is capable of
accelerating the metabolism of a range of
lipid-soluble drugs including CBZ, VPA, ESM,
anticoagulants, steroids and cyclosporin3*.
Because its metabolism is saturable, the drug
provides a target for mono-oxygenase inhibitors,
such as allopurinol, amiodarone, cimetidine,
imipramine and some sulphonamides. Protein-
binding displacement interactions with anti-
epileptic drugs are only clinically relevant when
there is concomitant enzyme inhibition, as is the
case with PHT and VPA39.
Valproic acid
The anticonvulsant properties of VPA were
recognized serendipitously in 1963, when it was
used by Pierre Eymard as a solvent for a number
of other compounds. It is now established as
effective over the complete range of seizure types,
with particular value in the idiopathic generalized
epilepsies4’. VPA’s mechanism of action is poorly
understood4’. It is currently believed to exert its
antiepileptic effects, at least in part, by a
mechanism similar to that of PHT and CBZ,
limiting sustained repetitive firing by use- and
voltage-dependent effect on Na+ channels4’.
However, VPA has many active metabolities
whose effects are not well characterized and,
given its wide spectrum of antiepileptic activity,
other cellular mechanisms are likely to be
involved.
The starting dose for adults and adolescents
should be 500mg daily for 1 or 2 weeks,
increasing in most patients to 500 mg twice daily.
Alterations thereafter can be made according to
the clinical status of the patient. Many patients
will need to take 1500-2000 mg daily. Since the
drug can take several weeks to become fully
effective, frequent dosage adjustments shortly
after initiating therapy are unwarranted. Because
VPA does not exhibit a clear-cut concentration
effect-toxicity relationship and daily variations in
concentration at a given dose are wide, routine
monitoring may not be helpful unless correlated
with the patient’s clinica situation. In addition,
patients frequently need and tolerate concentra-
tions up to 1040 pmolll (150 mg/l).
Establlshed antiepileptic drugs
Common unpleasant side-effects with VPA are
dose-related tremor, weight gain, thinning or loss
of hair (usually temporary), and menstrual
irregularities including amenorrhea. There is
some evidence implicating its use in the produc-
tion of polycystic ovariesJ3. Sedation is an
uncommon complaint although stupor and en-
cephalopathy can occur, possibly as a conse-
quence of underlying carnitine deficiency”.
Hepatoxicity, histologically a microvesicular
steatosis similar to that found in Reye’s syn-
drome, affects fewer than one in 20,000 treated
individuals. This appears to be a particular worry
in children under 3 years-of-age receiving anti-
epileptic polypharmacy, some of whom will have
a co-existent metabolic defect45. Hyperam-
monemia without hepatic damage can be dem-
onstrated in approximately 20% of all patients
receiving VPA46. This is usually transient, but
occasionally can present clinically with confusion,
nausea and vomiting, and clouding of conscious-
ness. Other sporadic problems include throm-
bocytopenia and pancreatitis.
VPA can inhibit a range of hepatic metabolic
processes, including oxidation, conjugation and
epoxidation4’. Targets include other antiepileptic
drugs, particularly PHT, PB, CBZ epoxide and
lamotrigine31. Aspirin displaces VPA from its
binding sites on plasma proteins and inhibits its
metabolism4’. VPA does not interfere with the
hormonal components of the oral contraceptive
pill.
Phenobarbital
The discovery of the antiepileptic properties of
PB was another classic example of serendipity. A
German physician, Alfred Hauptmann, had been
assigned rooms over his ward of patients with
epilepsy. Unable to get a night of uninterrupted
sleep, he provided them with a blanket prescrip-
tion of PB as an hypnotic. To his surprise and
delight, this treatment reduced the number of
seizures during the night and throughout the day.
PB is used worldwide and recent trials have
confirmed it to be as effective as CBZ and PHT in
abolishing partial and generalized tonic-clonic
seizures13. The anticonvulsant effects of PB are
believed to be mediated through potentiation of
gamma aminobutyric acid (GABA) inhibition by
binding to a specific site on the GABA* receptor
chloride channel complex. At a cellular level, PB
produces a prolongation of inhibitory postsynap-
tic potentials by increasing the mean chloride
channel opening time and the duration of
GABA-induced bursts4’.
165
PB is an easy drug to use clinically. To
minimize sedation, a low dose should be started
(60 mg in adults, 4 mg/kg in children), which can
be increased gradually according to clinical
requirements. The value of measuring circulating
PB is limited as the concentration associated with
optimal control varies considerablJ’. In addition,
the development of tolerance to its central
nervous side-effects makes the toxic threshold
imprecise. Nevertheless, an unexpectedly low or
high concentration may help to make the correct
clinical decision in an individual patient.
The major problem in the clinical use of PB lies
in its propensity to influence cognition, mood and
behaviour. It can produce fatigue, listlessness and
tiredness in adults and insomnia, hyperactivity
and aggression in children (and sometimes in the
elderly)“. Subtle impairment-of memory, mood
and learning capacity can occur in both groups.
Depression may be associated with long-term
administration, and arthritic changes and
Dupuytren’s contracture can be associated prob-
lems. Tolerance develops to the deleterious
cognitive effects but also, unfortunately, to its
anticonvulsant efficacy in some patient$l. PB is
the archetypal enzyme inducer and can accelerate
the metabolism of a long list of lipid-soluble
drugs5*.
Primidone
PRM was synthesized initially in an attempt to
produce a non-sedative barbiturate. The first
clinical trials in epilepsy took place in 1952. It was
not appreciated for some time that the parent
drug was biotransformed to PB and another
active metabolite, phenylethylmalonamide. The
efficacy of PRM is similar to that of PB, but it is
less well tolerated13. Not surprisingly, withdrawal
may result in seizure exacerbation. There is,
therefore, little to recommend this drug over PB
for those patients in whom treatment with a
barbiturate is contemplated. A poor correlation
exists between PRM concentrations and efficacy
and toxicit$‘. The major metabolite, PB, can be
measured if compliance is an issue.
Ethosuximide
ESM was the product of a successful search by
Parke Davis for a less toxic alternative to
trimethadione in the treatment of absence sei-
zures. It was introduced into clinical practice in
1958 and retains a limited place in the antiepilep-
tic armamentarium. ESM acts by reducing low
threshold, transient, voltage-dependent calcium
 166
conductance in thalamic neurons53. Slow intro-
duction is sensible to anticipate the development
of gastrointestinal and central nervous system
side-effects. In children over 6 years, 500 mg/day
is a reasonable starting dose, with further
increments as necessary to a maximum amount of
l-2g daily. For most patients the dose can be
increased every 2-4 weeks. Infants require higher
weight-related dosing, i.e. 15-30 mg/kg daily.
The side-effects of ESM are dose-related and
usually involve the gastrointestinal tract (nausea,
vomiting, abdominal pain) or central nervous
system (lethargy, dizziness, ataxia). They do not
readily correlate with high circulating ESM
concentrations and are usually volunteered by the
patient or his or her family. ESM dosage
adjustment can be assessed clinically or oc-
casionally electroencephalographically with tele-
metry. Routine drug monitoring is not necessary,
but a measurement can be helpful in distinguish-
ing a patient who is resistant to ESM’s phar-
macological effect from one who has low con-
centrations as a consequence of incomplete
compliance or rapid degradation33. ESM itself
does not interfere with drug metabolism, but does
provide a target for enzyme inducers such as PI-IT
and CBZ, or inhibitors such as VPA38.
Clobazam
Clobazam (CLB) was introduced as an anxiolytic
in 1975, and its anti-seizure activity was recog-
nized soon after. It is a 1,5 benzodiazepine that
exerts its activity by potentiating the inhibitory
action of gamma aminobutyric acid54. CLB is a
useful adjunctive drug in refractory epilepsy”.
Only a minority of patients, however, achieve a
worthwhile improvement in seizure control on
long-term dosing due to the development
tolerance. Some of these will have an anatomical
basis for their seizure disorde15. Nevertheless, a
useful proportion (lo-20%) will become seizure-
free”. There is some evidence that intermittent
use of CLB reduces the likelihood of tolerance56.
Short-term administration, e.g. 20-30 mg daily
for 3 days, can be effective in women with
catamenial exacerbations and as ‘cover’ for
special events such as holidays, weddings and
surgery. A single dose of 30mg can have a
prophylactic effect if taken immediately after the
fnst seizure in patients who suffer regular clusters
of complex partial or secondary generalized
seizures. CLB’s structure differs slightly from
those of clonazepam and diazepam, and this may
account for a lesser propensity to produce
sedation. Nevertheless, depression, irritability
M.J. Brodie 81 M.A. Dichter
and tiredness are commonly reported5’. As with
barbiturates, deterioration in behaviour and
mood .disturbance can occur, particularly in
patients with learning disabilities in whom CLB
should probably be avoided. Withdrawal seizures
can also be a problem.
Clonazepam
CZP has efficacy against absences, myoclonic
jerks and tonic-clonic seizures58. As with other
benzodiazepines, sedation and tolerance substan-
tially reduce its usefulness. Few patients respond
well to this drug, and nearly 50% will have an
exacerbation of their epilepsy when it is
withdrawn5’. Accordingly, it now has a narrow
role in the modem management of epilepsy,
possibly limited to refractory myoclonic seizures.
Like other benzodiazepines, CZP should only be
prescribed as a last resort for patients with
learning difficulties.
lNDlVlDUALlZATlON OF THERAPY
The choice of drug for the individual patient
depends on the seizure type or types and the
likely tolerability to potential side-effects.
Another important factor is the prescribing
clinician’s experience. For most patients with
partial and tonic-clonic seizures, CBZ is a good
first choice. Some individuals have early problems
with rash or central nervous system toxicity. If,
because of intermittent side-effects, the drug
needs to be administered more than twice daily,
compliance can become an issue. In such circum-
stances a controlled-release formulation should
of be substituted. CBZ induces the metabolism of
the hormonal components of the oral contracep-
tive, and this may cause difficulty in ensuring
adequate protection against unwanted pregnancy.
Its enzyme-inducing properties can be a major
drawback in patients receiving other lipid-soluble
drugs such as warfarin, cyclosporin, etc. The
association of CBZ with neural tube defects
should also be kept in mind when prescribing the
drug in female patients of child-bearing
potentialm.
PHT is as effective as CBZ against partial and
tonic-clonic seizures. It can be administered once
or twice daily and a useful parenteral preparation
is available. Gum hypertrophy, hirsutism and
changes in facial features are unpredictable
side-effects. Advice on mouth hygiene can pre-
vent the first of those, and the latter two are often
 Established antiepileptic drugs
mild and can be reversible. Teratogenesis is a
potential problem in this patient population, as is
PHT’s induction effect on the oral contraceptive
pill and other lipid soluble drugs. Saturation
kinetics make PI-IT a poor choice for patients
whose compliance is likely to be erratic. It is
usually necessary to monitor its circulating
concentration.
VPA is particularly effective for the idiopathic
generalized epilepsies, although it also has
efficacy against localization-related tonic-clonic
and partial seizures. Its tendency to cause weight
gain and patchy hair loss can cause problems,
although many patients tolerate the drug well.
Teratogenesis, particularly neural tube defects, is
undoubtedly a worq@ as is its reported effect in
producing polycystic ovaries43. These can be a
particular concern to young women with
idiopathic generalized seizures, in whom there is
. an expanding place for lamotrigine31. Care should
be taken when prescribing VPA for infants and
young children with suspected inborn errors of
metabolism because of the particular risk of
hepatotoxicity.
PB, which is as effective as CBZ and PI-IT
against partial seizures, is usually reserved for
patients intolerant to these agents, as it is sedative
and can produce depression. PB is an unattractive
alternative for children, because of its deleterious
influence on learning and behaviour. In addition,
it is the archetypal enzyme inducer and acceler-
ates the metabolism of a range of other lipid-
soluble drugs. On the positive side, PB is
inexpensive and easy to use. Accordingly, it still
has an important role to play in the management
of epilepsy in the third world, where cost is the
overwhelming consideration in the choice of
therapy.
POLYPHARMACY
More than 25% of people with epilepsy appear
refractory to optimal therapy with a single
antiepileptic drug. Many of these will have partial
seizures secondary to an underlying anatomical
abnormality. The temptation is to add other drugs
until acceptable improvement occurs. Current
evidence suggests, however, that this strategy will
be successful in only around 10% of such
patients 62. It is likely that combining established
drugs with similar, multiple and overlapping
modes of action is responsible for this disappoint-
ing outcome. Recent placebo-controlled studies
with newer agents, some of which have specific
mechanisms of action, hold out the possibility of
167
more promising combination therapy for refrac-
tory partial seizures’*.
Combining antiepileptic drugs results in a
number of deleterious adverse interactions
largely involving hepatic metabolic processes. No
fewer than four of the established drugs, CBZ,
PHT, PB and PRM (metabolized in part to PB)
have the ability to induce the synthesis of
oxidative and conjugating enzymes3g. All will
accelerate the breakdown of VPA, ESM and
CZP. When given together, mutual enzyme
induction is often the result. The exception
appears to be with CBZ and PHT. When
combined, PHT induces CBZ metabolism,
whereas CBZ, paradoxically, reduces the clear-
ance of PHp3. The result is a rise in PHT
concentration accompanied by a fall in that of
CBZ. Interactions with CBZ can be complicated
by an increase in the concentration of the active
metabolite, CBZ lo,11 epoxide, which occurs
during dual therapy with enzyme-inducing anti-
epileptic drugs and with the inhibitor VPAU.
VPA also inhibits PHT, PB and lamotrigine
metabolism65. The extent of the metabolic inter-
actions between antiepileptic drugs varies sub-
stantially among patients. With many, there will
be no clinical repercussions. Others will suffer a
major alteration in circulating concentration
resulting in seizure exacerbation or neurotoxicity.
Measuring the total drug concentration will often
help with appropriate dosage adjustment@. It is
important, too, that discontinuing an enzyme-
inducing or inhibiting drug may have a substantial
influence on the concentration of the remaining
agent@‘.
It has been known for some time that
antiepileptic drugs in combination can impair
cognition68. This may, in part, be due to the
severity of the underlying seizure disorder,
particularly in the presence of brain pathology6’.
However, the weight of evidence strongly sup-
ports an additive effect with established anti-
epileptic drugs and has contributed to the strong
pressure toward monotherapy”. A multifactorial
situation involving seizure severity, anatomical
disease, and kinetic and dynamic drug interac-
tions offers the most plausible explanation for this
phenomenon. Antiepileptic polypharmacy is also
more likely to result in teratogenesis and this will
be discussed later.
REFRACTORY EPILEPSY
As many as 20% of patients referred to a
specialist service with refractory seizures will not
have epilepsy at all. Most of these will have
 168
pseudoseizures71. Clues to this diagnosis include
wide variation in the clinical presentation of the
events, aggressive behaviour during the episodes,
normal or non-specific electroencephalographic
changes, and worsening of seizure control on
adding a small dose of an antiepileptic drug.
Some of these patients will be young women, who
have been sexually abused’*. Other patients do
have epilepsy, but will be receiving inappropriate
treatment, in particular CBZ or PHT for syn-
dromes involving myoclonic jerks. A few will
have absences treated with drugs other than VPA
or ESM. Some will have an indolent glioma.
Erratic compliance or covert alcoholism or drug
abuse are other reasons for apparent
‘refractoriness’.
Before treating a patient with more than one
antiepileptic drug, all reasonable monotherapy
options should be exhausted. This may consist of
just one first-line drug tolerated at high dosage in
some circumstances, e.g. if there is an underlying
resectable lesion. An alternative drug should be
introduced slowly. If successful, the original drug
can often be withdrawn, although not in all
patients 73. If seizure control remains poor at the
highest tolerated dose of a second first-line drug,
it is time to try dual therapy. There are no
controlled clinical trials to tell us what are the
best drugs to combine. For partial and general-
ized seizures, many specialists employ two of the
three first-line agents, namely CBZ, VPA and
PHT, although their mechanisms of action are
complex and overlapping. A more logical ap-
proach may be to add one of the newer
antiepileptic drugs with a different mechanism of
action. For myoclonic seizures not responding to
VPA, lamotrigine or a benzodiazepine such as
CZP can be added, while intractable typical or
atypical absences may respond to VPA and
ESM74 or lamotrigine75 in combination. It would
be counter-productive to introduce a sedative
drug such as PB or PI&l at this stage, since, if this
strategy is ineffective, subsequent withdrawal of
the barbiturate will almost certainly produce
withdrawal seizures. If the problem is refractory
partial seizures, the addition of one of the newer
drugs-lamotrigine, vigabatrin, gabapentin, ox-
carbazepine, topiramate or tiagabine-may be
more appropriate”. There is mounting evidence
to suggest that VPA and lamotrigine in combina-
tion may be particularly effective for a number of
seizure types3’,“.
In some patients truly refractory to phar-
macotherapy, the law of diminishing returns will
require patient and doctor to accept the persis-
tence of some seizures. In such a situation, it is
important to balance adequacy of control with
M.J. Brodie & MA. Dichter
optimal quality of life. If seizures continue despite
treatment with multiple antiepileptic drugs, little
can be lost in many patients by gradually reducing
the number of drugs and simplifying the dosage
schedules. This manoeuvre will often, paradoxi-
cally, reduce seizure frequency76. Unfortunately,
in a few patients seizure frequency may dramati-
cally worsen or tonic-clonic seizures emerge.
Slow reduction will avoid the likelihood of status
epilepticus. Producing less intrusive episodes,
abolishing tonic-clonic seizures, preventing falls,
and decreasing automatisms can all be acceptable
end-points in patients with difficult epilepsy.
Other important options include early work-up
for epilepsy surgery or trying a range of new
antiepileptic drugs. This latter alternative is
unlikely to be successful if the patient is already
taking high doses of three antiepileptic drugs!
THERAPEUTIC DRUG MONITORING
Although monitoring antiepileptic drug con-
centrations can be helpful in optimizing the dose
in some patients, rigorous adherence to a narrow
concentration range is likely to result in the
doctor treating the drug level and not the
patient”. The ‘therapeutic’ or, better, ‘target’
range can only be regarded as an approximation,
based as it is on population data. Some patients
will do well with levels below the lower limit of
the target’*, whereas other require concentrations
above the range to become seizure-free79. All too
often the dose is reduced because of a reported
‘toxic’ level, when the patient is symptom- and
seizure-free. In other patients who are not fully
controlled, the dose may not be increased
appropriately for fear of stepping outside a
mythical ‘therapeutic’ range.
Nevertheless, appropriate use of concentration
monitoring, particularly on site at an epilepsy
clinic, can promote an improved outcomexO. Once
treatment is initiated, the goal is complete seizure
control using a single drug with no or minimal
side-effects. It makes sense, therefore, to ensure
that the concentration in a patient with newly
diagnosed epilepsy, who may well have infre-
quent seizures, is within the relevant target range
to optimize the likelihood of perfect control.
Levels are useful also in encouraging compliance.
Monitoring can be helpful in a well-controlled
patient with a break through seizure. A lower
than expected concentration may indicate a
change in metabolism (e.g. CBZ inducing its own
metabolism), may reflect an unexpected interac-
tion with another drug, or may suggest a lapse in
compliance, a particular problem in adolescents.
Levels can distinguish between too much
Established antiepileptic drugs
medication and too little. For example, some
patients with partial seizures complain of memory
difficulties. This can result from the pathology
underlying the seizure disorder, from the seizures
themselves, from subclinical ictal discharges, or
from a deleterious effect of medication. A the general population compared with 3% for
concentration measurement can determine
whether or not a new symptom is likely to
represent a side-effect of treatment. Monitoring
can also be useful in circumstances where
concentrations vary unexpectedly or unpredic-
tably. For a drug like PHT, which undergoes
saturation kinetics, neurotoxic symptoms may
occur after a small increment in dose. Concentra-
tions may be altered by co-administered medica-
tions (especially enzyme inducers and inhibitors),
during pregnancy, in patients with renal or
hepatic impairment, or when other medical
conditions, e.g. thyroid disorders, co-exist.
Although therapeutic drug monitoring can facilit-
ate patient management, its major limitations
must also be appreciated”. Drug concentrations
vary widely over the course of a day and so
monitoring is best related to the time since
dosing. Another potential pitfall is the presence
of active metabolites (e.g. with VPA or CBZ).
Certian drug interactions will lower the parent
drug level, while elevating that of a potentially
toxic metabolite.
Finally, the therapeutic and unwanted effects of
antiepileptic drugs may occasionally be reflected
better by free than by total concentrationss2. In
most circumstances, these will correlate well. This
is not the case, however, when protein-binding is
reduced. Such situations occur in the neonate, in
old age, in pregnancy, in severe renal and hepatic
impairment, and with the co-prescription of drugs
that can alter binding. The methodology for
measuring free levels is more complex, less
reliable, and more expensive than assaying total
drug concentrations. However, when a patient
experiences an unexpected degree of toxicity at
apparently modest or moderate dosage, measure-
ment of the free level may indicate the cause.
PREGNANCY
Most women with treated epilepsy undergo
uneventful pregnancies and deliver healthy
babiess3. During pregnancy, however, metabolic
and excitatory processes change and closer
attention should be given to antiepileptic drug
concentrations. Total levels of some established
drugs will fall, particularly those of PHP.
Women who are well controlled will usually
remain so during pregnancy and delivery. How-
169
ever, those who continue to report seizures may
suffer a deterioration. There is a small increased
incidence of significant fetal malformations in the
offspring of epileptic women, even if they are
untreated&. Commonly quoted figures are 2% in
women with epilepsy. The risk increases with the
number of antiepileptic drugs, being approxim-
ately 3% for one drug (similar to background),
5% for two, 10% for three, and over 20% in
women taking as many as four antiepileptic
drugs 86. The mechanism is likely to involve the
production of mutagenic metabolitess7.
A syndrome consisting of facial dimorphism,
cleft lip and palate, cardiac defects, digital
hypoplasia and nail dysplasia has been identified.
This was initially ascribed to hydantoins including
PHT (‘fetal-hydantoin syndrome’), but is now
known to occur with most other antiepileptic
drugs, including CBZ and VPAs7. There are no
clear data indicating differences in safety among
PHT, PB or PRM. Current evidence suggests,
however, that VPA and CBZ are associated with
a low incidence of neural tube defects, probably
l-2% and OS-l%, respectively88.
Although it would be ideal for a woman
contemplating pregnancy to have her antiepilep-
tic medication withdrawn, for many this would
result in recurrence or exacerbation of seizures,
which could be dangerous for both mother and
baby. If the criteria for discontinuation of
medication are met, this should be done over a
suitable interval before conception. Other pati-
ents should be tapered to a minimal effective dose
of, if possible, a single antiepileptic drug. In
addition, supplemental folic acid should be
administered preconceptually in an attempt at
preventing neural tube defects. To be effective,
folate must be taken during the first 5 weeks of
gestation. The current recommendations are
5 mg/day folic acid for women who have had a
child with such a defect, and 0.4 mg daily in other
women planning a pregnancy”. It seems sensible,
however, to recommend the full 5 mg folate daily
for all women receiving treatment with anti-
epileptic drugs=.
Eznyme-inducing antiepileptic drugs (CBZ,
PHT, PB and PRM) can cause transient and
reversible deficiency in vitamin K-dependent
clotting factors in the neonate. Following a
traumatic birth, there is an increased risk of
intracerebral hemorrhage. Accordingly, pregnant
women receiving one or more of these drugs
should be treated with 20 mg of vitamin K1 daily
during the last month of pregnancy, and babies at
risk should receive a single intramuscular dose of
1 mg of vitamin K, immediately after birthgO.
170
WITHDRAWING TREATMENT
A successful outcome for treated epilepsy can be
regarded as a patient who becomes seizure-free in
the absence of, adverse effects. Such individuals
usually lead integrated lives, with successful
intellectual and emotional development and
positive educational and vocational achieve-
ments. Indeed, many such patients can eventually
have their medication withdrawn and will remain
seizure-free”. Patients who are ‘doing well’ may
want to stop treatment for a variety of reasons,
including the presence of unpleasant side-effects
or the subjective perception of subtle deteriora-
tion in cognitive function. In addition, taking
medication does not equate with normal health. If
drug therapy is stopped and seizures do not recur,
there is a sense of being ‘cured’ of the stigma of
epilepsy. Finally, patients may want to become
parents and are often concerned about the
possible negative effects of antiepileptic drugs on
reproductive function and about the spectre of
teratogenesis. Several studies have shown that
after a period of optimal seizure control, medica-
tion can be stopped without recurrence (at least
within several years) in more than 60% of
patients92-96.
There are no data to indicate an optimum
length for the seizure-free period; most studies
used 2-5 years9’. Seizure type or epilepsy
syndrome is not absolutely predictive of recur-
rence. However, a few specific childhood syn-
dromes, such as benign epilepsy of childhood with
rolandic spikes and benign familial neonatal
convulsions, tend to do well after drug with-
drawal, whereas juvenile myoclonic epilepsy
conveys a high probability of relapse. Some forms
of idiopathic generalized seizures, either absence
or tonic-clonic, if existing alone, are less likely to
recur after control is achieved. However, even
complex partial seizures can disappear after a
long period of perfect control.
Individuals with the highest probability of
remaining seizure-free after medication is discon-
tinued are those reporting no seizures for a long
period, those with relatively few seizures before
control was achieved, those taking a single
antiepileptic drug, and those with a normal
neurological examination with no structural brain
lesion98. The role of the EEG in predicting
seizure recurrence is somewhat more
controversialWV9*. Some studies have indicated a
low predictive value for this investigation,
whereas others have suggested a better prognosis
for continued remission when the EEG is normal.
There are no protocols defining optimal regimens
M.J. Brodie & M.A. Dichter
for tapering medication. Most specialists advise
slow reduction over periods ranging from 2 to 6
months. If the patient is taking antiepileptic
polypharmacy, one drug should be completely
withdrawn before the second is tapered. More
than 90% of recurrences will occur during the 1st
year following withdrawal, and most will present
during the taper period or shortly after.
EPILEPSY PROPHYLAXIS
Some individuals are known to be at high risk of
developing epilepsy. These include those with
severe head trauma, significant central nervous
system infection, supratentorial brain tumour,
ventricular shunt, arteriovenous malformation,
haemorrhagic stroke, etc. There are few data
exploring the efficacy of antiepileptic drugs in
preventing its development. One controlled trial
showed that PHT could suppress seizures acutely
after head injury99, but another did notIm. There
is doubt, too, regarding their efficacy in prevent-
ing the development of chronic epilepsy10’*‘02.
Other trials have suggested that high serum
antiepileptic drug concentrations are needed for
successful prophylaxis’03. At the present time, it is
difficult to recommend prophylactic antiepileptic
treatment, but more research is needed in this
important area.
REFERRAL TO AN EPILEPSY CENTRE
Some individuals with uncomplicated epilepsy are
cared for by family practitioners, general physici-
ans or pediatricians. However, as proper diag-
nosis, accurate seizure and epilepsy syndrome
classification, and appropriate selection and super-
vision of medication is essential for optimal
management, is is often in the patient’s best
interests to involve a neurologist or other epilepsy
specialist early in the evaluation. Patients with
complicated seizure problems are more appropri-
ately cared for by an epileptologist, who can
provide additional expertise in diagnosis and
treatment.
If patients continue to have seizures, not clearly
due to erratic compliance with medication,
referral to an epilepsy centre can be essential for
a number of reasons. Firstly, it appears that the
more seizures an individual experiences, the more
difficult they are to control. That ‘seizures (may)
beget seizures’ has been hypothesized for more
than 100 years’03. Thus, tolerating prolonged
periods of uncontrolled seizures may be prognos-
tically deleterious. Secondly, seizure classification
may be hard to determine based on untrained
Established antiepileptic drugs
witness accounts. Video-EEG monitoring can
provide a definitive diagnosis and classification,
which can lead to more effective therapy. Some
patients with ‘uncontrolled seizures’ are really
having pseudoseizures, which are better treated
by psychotherapy and, often, reduction in or
withdrawal of antiepileptic medication. Thirdly,
patients with uncontrolled seizures often receive
polypharmacy and this may produce significant
toxicity, which can be reduced by careful phar-
macological simplification under monitored
conditions’“4. Fourthly, patients who have failed
to respond adequately to the standard drugs will
have the opportunity to try one or more newer
agents under carefully controlled conditions at an
epilepsy centre. Lastly, appropriate decisions
about surgery can best be made in a specialized
centre.
CONCLUSIONS
In this paper, we have attempted an overview of
the practical management of epilepsy using
established antiepileptic drugs. When prescribing
these agents, the following 12 basic rules should
be followed:
1. Ensure that the diagnosis of epilepsy is
soundly based, that there is no other under-
lying correctable condition, and that the
patient understands and accepts the necessity
of taking antiepileptic drug therapy for some
years;
2. Choose the best drug for each individual
patient using the international classifications
of seizure types and epilepsy syndromes;
3. Start at low dosage and titrate upwards in
increments to a maintenance amount that
controls the seizures but permits tolerance
to central nervous system side-effects and
avoids concentration-dependent toxicity;
4. Keep the regimen simple with once or
twice-daily dosing when possible;
5. Counsel the patient early regarding the
implications of the diagnosis, the prophylac-
tic nature of the therapy, and the importance
of perfect compliance;
6. In patients proving resistant to a first-line
drug but not complaining of side-effects,
titrate the dose to the limit of tolerability
without worrying about the circulating
concentration:
7. As monotherapy is the preferred approach
for the majority of patients, a second drug
should usually be substituted before embark-
ing upon dual therapy;
171
8. It may more logical to add in a new drug with
a potentially complementary mechanism of
action than to combine established agents;
9. Simplify dosage schedules and drug regimens
as much as possible in patients who are truly
refractory to antiepileptic polypharmacy;
10. In refractory epilepsy, aim for the best
seizure control consistent with optimal qual-
ity of life;
11. Refer patients to an epilepsy centre when
seizures persist despite treatment with first-
line drugs or in patients with problems
tolerating the established agents;
12. Consider epilepsy surgery early for patients
with drug-resistant partial seizures or gene-
ralized seizures with multiple falls.
Good communication with the patient and his or
her family, together with simple, clear and
sympathetic explanations of the treatment plan
and potential pitfalls, is essential throughout.
Counselling about the implications of the diag-
nosis is an important part of optimal
management.
REFERENCES
.1. Shorvon, SD. Epidemiology, classification, natural
history and genetics of epilepsy. Loncef 1990; 336: 93-96.
2. Hauser, W.A., Annegers, J.F. and Kurland, L.J.
Prevalence of epilepsy in Rochester, Minnesota: 1940-
1980. Epilepsiu 1991; 32: 429-445.
3. Watts, A.E. The natural history of untreated epilepsy in
a rural community in Africa. Epilepsia 1992; 33:
464-468.
4. Reynolds, E.H. Changing view of prognosis of epilepsy.
British Medical Journal 1990; 301: 1112-1114.
5. Camfield, C., Camfield, P., Gordon, K. and Dooley, J.
Does the number of seizures before treatment influence
ease of control or remission of childhood epilepsy? Not
if the number is 10 or less. Neurology 1996; 46: 41-44.
Chadwick, D.W. Diagnosis of epilepsy. Lancer 1990; 336:
291-295.
Brodie, M.J. and Dichter, M.A. Antiepileptic drugs.
New England Journal of Medicine 1996; 3341 168-175.
Leestma, J., Walczak, T., Hughes, J., KaleIkar, M., Teas,
S. A prospective study on sudden unexpected death in
epilepsy. Annals of Neurology 1989; 26: 195-203.
9. Lip, G.Y. and Brodie, M.J. Sudden unexpected death in
epilepsy: an avoidable outcome? Journal of the Royal
Society of Medicine 1992; 85: 609-611.
10. Hauser, W., Rich, S., Annegers, J. and Anderson, V.
Seizure recurrence after a hrst unprovoked seizure: an
extended follow-up. Neurology 1990; 40: 1163-1170.
11. Shinnar, S., Berg, A. and Moshe, S. Risk of seizure
recurrence following a first unprovoked seizure in
childhood: a prospective study. Pediatrics 1990; 85:
1076-1085.
12. Gilad, R., Lampl, Y., Gabbay, U., Eshely, Y. and
Sarova-Pinhab, I. Early treatment of a single generated
tonic-clonic seizure to prevent recurrence. Archiues of
Neurology 1996; 53: 1149-1152.
172 M.J. Brodie & M.A. Dichter

13. Mattson, R.H., Cramer, J.A., Collins, J.F. et al. Journal of Neurology, Neurosurgery and Psychiany
Comparison of carbamazepine, phenobarbital, 1984; 47: 642-644.
phenytoin and primidone in partial and secondary 29. Brodie, M.J. Drug interactions and epilepsy: Epilepsia
generalised tonic-clonic seizures. New England Journal 1992; 33 (Suppl. 1): S13-S22.
ofMedicine 1985; 313: 145-151. 30. Mattson, R.H., Cramer, J.A., Damey, P.D. and Naf-
14. Mattson, R.H., Cramer, J.A. and Collins, J.F. A tolin, F. Use of oral contraceptives by women with
comparison of valproate with carbamazepine for the epilepsy. Journal of the American Medical Association
treatment of complex partial seizures and secondarily 1986; 256: 238-240.
generalised tonic-clonic seizures. New England Journal 31. Brodie, M.J. Lamotrigine-an update. Cunadinn Journal
of Medicine 1992; 327: 765-771. of Neurological Science 1996; 23 (Suppl. 2): S6-S9.
15. Richens, A., Davidson, D.L.W., Cartlidge, N.E.F. and 32. Shukla, S., Godwin, C.D., Long, L.E.B. and Miller,
Easter, D.J., on behalf of the adult EPITEG collabora- M.G. Lithium-carbamazepine neurotoxicity and risk
tion group. A multicentre comparative trial of sodium factors. American Journal of Psych&y 1984; 141:
valproate and carbamazepine in adult onset epilepsy. 1604-1606.
Journal of Neurology, Neurosurgery and Psychiatry 33. Brodie, M.J. and Feely, J. Practical clinical pharmacol-
1994; 57: 682-687. ogy. Therapeutic drug monitoring and clinical trials.
16. Verity, C.M., Hosking, G. and Easter, D.J., On behalf of British Medical Journal 1988: 296: 1110-l 114.
the paediatric EPITEG collaboration group. A multi- 34. Selzer, M.E. The action of phenytoin on a composite
centre comparative trial of sodium valproate and electrical-chemical synapse in the lamprey spinal cord.
carbamazepine in paediatric epilepsy. Deuelopmenral Annals of Neurology 1978; 3: 202-206.
Medicine and Child Neurology 1995; 37: 97-108. 35. Yaari, Y., Pincus, J.H. and Argov, Z. Depression of
17. De Silva, M., McArdle, B., McGowan, M. et al. synaptic transmission by diphenylhydantoin. Ann& of
Randomised comparative monotherapy trial of pheno- Neurology 1977; 1: 334-338.
barbitone, phenytoin, carbamazepine, or sodium valpro- 36. Thomson, A.H. and Brodie, M.J. Pharmacokinetic
ate for newly diagnosed childhood epilepsy. Lancer 1996; optimisation of anticonvulsant therapy. Clinical Phar-
347: 709-713. macokinetics 1992: 23: 216-230.
18. Dichter, M.A. and Brodie, M.J. New antiepileptic drugs. 37. Botez, MI., Ezzedine, A. and Vezina, J.L. Cerebellar
New England Journal of Medicine 1996; 334: 1.583-1590. atrophy in epileptic patients. Canadian Journal of
19. Editorial. Carbamazepine update. Lancer 1989; ii: Neurological Science 1988; 15: 299-303.
595-597. 38. Pisani, F., Perucca, E. and Di Perri, R. Clinically
20. Liporace, J., Sperling, M. and Dichter, M. Absence relevant antiepileptic drug interactions. Journal of
seizures and carbamazepine in adults. Epilepsia 1994; 35: International Medical Research 1990; 18: 1-15.
1026-1028. 39. McInnes, G.T. and Brodie, M.J. Drug interactions that
21. McLean, M.J. and Macdonald, R.L. Carbamazepine and matter-a critical reappraisal. Drugs 1988; 36: 83-110.
10,ll epoxycarbamazepine produce use- and voltage- 40. Editorial. Sodium valproate. Lancer 1988; II: 1229-1231.
dependent limitation of rapid firing of action potentials 41. Macdonald, R. and Kelly, K. Antiepileptic drug mecha-
of mouse central neurons in cell culture. Journal of nisms of action. Epilepsiu 1993; 34 (Suppl. 5): l-8.
Experimental Pharmacology and Therapeutics 1986; 238: 42. McLean, M.J. and Macdonald, R.L. Sodium valproate,
727-738. but not ethosuximide, produces use- and voltage-
22. Dichter, M.A. Old and new mechanisms of antiepileptic dependent limitation of high frequency repetitive firing
drug actions. Epilepsy Research 1993; 10 (Suppl.): 9-17. of action potentials of mouse central neurons in cell
23. Larkin, J.G., McKee, P.J.W. and Brodie, M.J. Tolerance culture. Journal of Pharmacology and Experimental
to psychomotor side-effects with carbamazepine. British Therapeutics 1986; 237: 1001-1011.
Journal of Clinical Pharmacology 1992; 33: 111-114. 43. Isojarvi, J.L., Laatikainen, T.J., Pakarinen, A.J., Jun-
24. Macphee, G.J.A., Butler, E. and Brodie, M.J. Intradose tunen, K.T. and Myllyla, V.V. Polycystic ovaries and
and circadian variation in circulating carbamazepine and hyperandrogenism in women taking valproate for
its epoxide in epileptic patients: a consequence of epilepsy. New England Journal of Medicine 1993; 329:
autoinduction of metabolism. Epilepsiu 1987; 28: 286-294. 1383-1388.
25. Gillham, R.A., Williams, N., Weidmann, K., Butler, E., 44. Triggs, W.J., Bohan, T.P., Lin, S. and Willmore, J.
Larkin, J.G. and Brodie, M.J. Concentration-effect Valproate-induced coma with ketosis and carnitine
relationship with carbamazepine and its epoxide on insufficiency. Archives of Neurology 1990; 47: 1131-
psychomotor and cognitive function in epileptic patients. 1133.
Journal of Neurology, Neurosurgery and Psychiatry 45. Dreifuss, F.F., Santilli, N., Langer, D.H., Sweeney, K.P.,
1988; 51: 929-933. Moline, B.A. and Menander, K.B. Valproic acid hepatic
26. McKee, P.J.W., Blacklaw, J., Carswell, A., Gillham, fatalities: a retrospective review. Neurology 1987; 37:
R.A. and Brodie, M.J. Double-dummy comparison 379-385.
between once and twice daily dosing with modified- 46. Patsalos, P.N., Wilson, S.J., Popovik, M., Cowan,
release carbamazepine in epileptic patients. British J.M.A., Shorvon, S.D. and Hjelm, M. The prevalence of
Journal of Clinical Pharmacology 1993; 36: 257-261. valproic acid associated hyperammonaemia in patients
27. McKee, P.J.W., Blacklaw, J., Butler, E., Gillham, R.A. with intractable epilepsy resident at the Chalfont Centre
and Brodie, M.J. Monotherapy with conventional and for Epilepsy. Journal of Epilepsy 1993; 6: 228-232.
controlled-release carbamazepine: a double-blind, 47. Levy, R.H. and Koch, K.M. Drug interactions with
double dummy comparison in epileptic patients. Brifish valproic acid. Drugs 1982; 24: 543-556.
Journal of Clinical Pharmacology 1991; 32: 99-104. 48. Goulden, K.J., Dooley, J.M., Camfield, P.R. and Fraser,
28. Chadwick, D.W., Shaw, M.D.M., Foy, P., Rawlins, M.D. A.D. Clinical valproate toxicity induced by acetylsali-
and Turnbull, D.M. Serum anticonvulsant concentra- cylic acid. Neurology 1987; 37: 1392-1394.
tions and the risk of drug-induced skin eruptions. 49. Twyman, R.E., Rogers, C.J. and Macdonald, R.L.
Established antiepileptic drugs 173

Differential regulation of gamma-aminobutyric acid 68. Kalviainen, R., Aikia, M. and Riekkinen, P.J. Cognitive
receptor channels by diazepam and phenobarbital. adverse effects of antiepileptic drugs. CNS Drugs 1996;
Annals of Neurology 1989; 25: 213-220. 5: 358-368.
50. Theodore, W.H. Rational use of antiepileptic drug 69. Beran, R.G. and Flanagan, P.J. Psychyosocial sequelae
levels. Pharmacology and Therapeutics 1992; 54: 297- of epilepsy: the role of associated cerebral pathology.
305. Epilepsia 1987; 28: 107-110.
51. Pritchard, J.W. and Mattson, R.A. Barbiturates: an 70. Brodie, M.J., McPhail, E., Macphee, G.J.A., Larkin,
update. In: Recent Advances in Epilepsy 3 (Eds T.A. J.G. and Gray, J.M.B. Psychomotor impairment and
Pedley and B.S. Meldrum). Edinburgh, Churchill anticonvulsant therapy in adult epileptic patients.
Livingstone, 1986; pp. 261-277. European Journal of Clinical Pharmacology, 1987; 31:
52. Patsalos, P.N. and Duncan, J.S. Antiepileptic drugs. A 655-660.
review of clinically signiliant drug interactions. Drug 71. Ozkara, C. and Dreifuss, F.E. Differential diagnosis in
Safety 1993; 9: 156-184. pseudoepileptic seizures. Epilepsia 1993; 34: 294-298.
53. Coulter, D.A., Huguenard, J.R. and Prince, D.A. 72. Alper, K., Devinsky, O., Penine, K., Vazquez, B. and
Characterization of ethosuximide reduction of low- Luciano, D. Non-epileptic seizures and childhood sexual
threshold calcium current in thalamic neurons. Annals of and physical abuse. Neurology 1993; 43: 1950-1953.
Neurology 1989; 25: 582-593. 73. Brodie, M.J. and Roy, K.B. One drug or two? A
54. Rogawski, M.A. and Porter, R.J. Antiepileptic drugs: double-blind comparison of adjuvant vigabatrin and
pharmacological mechanisms and clinical efficacy with valproate in carbamazepine-resistant epilepsy. Epilepsia
consideration of promising developmental stage com- 1996; 37 (Suppl. 5): 120.
pounds. Pharmacological Reviews 1990; 42: 223-286. 74. Rowan, A.J., Meijer, J.W.A., De Beer-Pawlikowski, N.,
55. Heller, A.J., Ring, H.A. and Reynolds, E.H. Factors van der Geest, P. and Meinardi, H. Valproate-
relating to dramatic response to clobazam therapy in ethosuximide combination therapy for refractory ab-
refractory epilepsy. Epilepsy Research 1988; 2: 276- sence seizures. Archives of Neurology 1983; 40: 797-802.
280. 75. Brodie, M.J., Yuen, A.W.C. and the 105 study group.
56. Feely, M. and Gibson, J. Intermittent clobazam for Lamotrigine substitution study: evidence for synergism
catamenital epilepsy: tolerance avoided. Journal of with sodium valproate? Epilepsy Research 1997 26:
Neurology, Neurosurgery and Psychiatry 1984; 27: 423-432.
1279-1282. 76. Theodore, W.H. and Porter, R.J. Removal of sedative,
57. Robertson, M.M. Current status of the 1,4 and hypnotic antiepileptic drugs from the regimes of patients
1,5-benzodiazepines in the treatment of epilepsy: the with intractable epilepsy. Annals of Neurology 1983; W:
place of clobazam. Epilepsia 1986; 27 (Suppl. 1): 320-324.
S27-S41. 77. Dodson, W.E. Level off. Neurology 1989; 39: 1009-1010.
58. Naito, H., Wachi, M. and Nishida, M. Clinical effects and 78. Woo, E., Chan, Y.M., Yu, Y.L. and Huang, G.Y. If a
plasma concentrations of long-term clonazepam mono- well stabilised patient has a subtherapeutic antiepileptic
therapy in previously untreated epileptics. Acta Neurol- drug level, should the dose be increased? A randomised
ogica Scandinavica 1987; 76: 58-63. prospective study. Epilepsia 1988; 29: 129-139.
59. Specht, U., Boenigk, H.E. and Wolf, P. Discontinuation 79. Gannaway, D.J. and Mawer, G.E. Serum phenytoin
of clonazepam after long-term treatment. Epilepsia concentration and clinical response in patients with
1989; 30: 458-463. epilepsy. British Journal of Clinical Pharmacology 1981;
60. Editorial. Teratogenesis with carbamazepine. Lancer 2: 833-839.
1991: 337: 1316-1317. 80. McKee, P.J.W., Percy-Robb, I. and Brodie, M.J.
61. Brodie, M.J. Management of epilepsy during pregnancy Therapeutic drug monitoring improves seizure control
and lactation. Lancer 1990; 336: 426-427. and reduces anticonvulsant side-effects in patients with
62. Schmidt, D. and Gram, L. Monotherapy versus poly- refractory epilepsy. Seizure 1992, 1: 275-279.
therapy in epilepsy. A reappraisal. CNS Drugs 1995; 3: 81. Commission on Antiepileptic Drugs, International
194-208. League Against Epilepsy. Guidelines for therapeutic
63. Zielinski, J.J. and Haidukewych, D. Dual effect of monitoring of antiepileptic drugs. Epilepsia 1993; 34:
carbamazepine-phenytoin interaction. Therapeutic Drug 585-587.
Monitoring 1987; 9: 21-23. 82. McKee, P.J.W. and Brodie, M.J. Therapeutic drug
64. Brodie, M.J., ,Forrest, G. and Rapeport, W.G. Car- monitoring. In: Epilepsy: A Comprehensive Textbook
bamazepine 10,ll epoxide concentrations in epileptics (Eds J. Engel and T.A. Pedley). New York, Raven Press
on carbamazepine alone and in combination with other (in press).
anticonvulsants. Britirh Journal of Clinical Pharmacol- 83. Delgado-Escueta, A. and Janz, D. Consensus guidelines:
ogy 1983; 16:747-750. preconception counselling, management and care of the
65. McKee, P.J.W. and Brodie, M.J. Pharmacokinetic pregnant woman with epilepsy. Neurology 1992; 42
interactions with antiepileptic drugs. In: New Anti- (Suppl. 5): 149-160.
convulsants: Advances in the Treatment of Epilepsy (Ed. 84. Brodie, M.J. Management of epilepsy during pregnancy
M.R. Trimble). Chichester, John Wiley, 1994: pp. l-33. and lactation. Lancer 1990; 336: 426-427.
66. McKee, P.J.W., Larkin, J.G., Brodie, A., Percy-Robb I. 85. Nakane, Y., Okuma, T., Takahashi, R. et al. Multi-
and Brodie, M.J. Five years of anticonvulsant monitor- institutional study on the teratogenicity and fetal toxicity
ing on site at the epilepsy clinic. Therapeutic Drug of antiepileptic drugs: a report of a collaborative study
Monitoring 1993; 15:83-90. group in Japan. Epilepsia 1980; 21: 663-680.
67. Duncan, J.S., Patsalos, P. and Shorvon, S.D. Effect of 86. Lindhout, D., Hoppener, R.S.E.A. and Meinard, H.
discontinuation of phenytoin, carbamazepine and Teratogenicity of antiepileptic drug combinations with
valproate on concomitant antiepileptic medication. special emphasis on epoxidation (of carbamazepine).
Epilepsia 1991; 32: 101-115. Epilepsia 1984.25: 77-83.
174
87. Yerby, M.S. and Leppik, I. Epilepsy and the outcomes of
pregnancy. Journal of Epilepsy 1990,3: 193-199.
88. Guthrie, E. and Brodie, M.J. Epilepsy, anticonvulsants
and pregnancy. In: Proceedings of XVrh European
Congress of Perinaral Medicine (Ed. F. Cockbum).
Camforth, Parthenon Publishing (in press).
89. Department of Health. Folic acid and the prevention of
neural tube defects: report from an expert advisory
group. DOH Health Publication Unit, Heywood, Eng-
land; 1992, pp. 21.
90. Zipursky, A. Vitamin K at birth. British Medical Journal
1996; 3l3: 179-180.
91. Editorial. Anticonvulsant drug withdrawal-hawks or
doves? Lancet 1991; 337: 1193-1194.
92. Shbmar, S., Vining, E., Mellits, E. er al. Discontinuing
antiepileptic medication in children with epilepsy after
two years without seizures. New England Journal of
Medicine 1985; 3W: 976-981.
93. Dean, J.C. and Penry, J.K. Discontinuation of anti-
epileptic drugs. In: A&epileptic Drugs (third edition)
(Eds R. Levy, R. Mattson, B. Meldrum and F. Dreifuss).
New York, Raven Press; 1989, pp. 133-142.
94. Medical Research Council Antiepileptic Drug Research
Group. Randomised study of antiepileptic drug with-
drawal in patients in remission. Lancer 1991; 337:
117.5-1180.
95. Mastropaolo, C., Tondi, M., Carboni, F., Manta, S. and
Zoroddu, F. Prognosis after therapy discontinuation in
children with epilepsy. European Neurology 1992; 32:
M.J. Brodle & M.A. Dichter
141-145.
96. Shinnar, S., Berg, A.T., Moshe, S.L. er al. Discontinuing
antiepileptic drugs in children with epilepsy: a prospec-
tive study. Annals of Neurology 1994; 354: 534-545.
97. Chadwick, D. A practical guide to discontinuing
antiepileptic drugs. CNS Drugs 1994; 2: 423-428.
98. Shinnar, S. and Berg, A.T. Withdrawal of antiepileptic
drugs. Current Opinion in Neuroiogy 1995; 8: 103-
106.
99. Young, B., Rapp, R., Norton, .I., Haack, D., Tibbs, P.
and Bean, J. Failure of prophylactically administered
phenytoin to prevent early post-traumatic seizures.
Journal of Neurosurgery 1983; 58: 236-241.
100. Young, B., Rapp, R., Norton, J., Haack, D., Tibhs, P.
and Bean, J. Failure of prophylactically administered
phenytoin to prevent late post-traumatic seizures.
Journal of Neurosurgery 1983; 58: 231-235.
101. North, J., PenhaU, R., Hanieh, A., Frewin, D. and
Taylor, W. Phenytoin and post-operative epilepsy.
Journal of Neurosurgery 1983; 58: 672-677.
102. Wohns, R. and Wyler, A. Prophylactic phenytoin in
severe head injuries. Journal of Neurosurgery 1979; 51:
507-509.
103. Gowers, W.R. Epilepsy and Other Chronic Convulsive
Diseases. London, Churchill, 1881.
104. Theodore, W., Schulman, E. and Porter, R. Intractable
seizures: long-term follow-up after prolonged in-patient
treatment in an epilepsy unit. Epilepsia 1983; 24:
336-343.