CHOOSING AMONG

ANTIEPILEPTIC DRUGS
Nathan B. Fountain

ABSTRACT
Selecting an antiepileptic drug (AED) for treatment of seizures is daunting because
there are now 20 from which to choose. Following simple principles allows a systematic approach to drug selection. Efficacy studies provide limited information. For
initial monotherapy of partial seizures, high-level evidence exists for the efficacy of
lamotrigine, carbamazepine, oxcarbazepine, and topiramate. For initial monotherapy of generalized seizures, high-level evidence is available for valproate, lamotrigine,
topiramate, and oxcarbazepine. For initial monotherapy of absence seizures, highlevel evidence exists for valproate, lamotrigine, and ethosaximide. All secondgeneration AEDs have efficacy as adjunctive therapy for partial seizures. AEDs are
often useful for comorbid conditions or have properties that should be avoided in
some groups. Thus, AEDs should usually be selected on the basis of comorbid
conditions, including depression, migraine, chronic pain, obesity, and nephrolithiasis, or patient characteristics, especially for women of childbearing potential and
older adults.

KEY POINT

The first principle

of prescribing
an antiepileptic
drug (AED) is
to select the
most efficacious
AED for the
seizure type
or epilepsy
syndrome
present.

Continuum Lifelong Learning Neurol 2010;16(3):121135.

GENERAL PRINCIPLES
Selecting among antiepileptic drugs
(AEDs) is intimidating because there are
now more than 20 AEDs from which to
choose (Table 6-1). However, use of a
few basic principles allows a simple
systematic approach to selecting the
most appropriate drug. This is facilitated greatly by the fact that most
second-generation AEDs have simple
pharmacokinetics and few drug interactions compared to the complexity of
conventional AEDs. The foundation of
selecting a proper drug is correctly diagnosing the problem through a careful history and evaluation to ensure
that the spells in question are seizures
and to characterize the seizure type
properly and determine the etiology
or epilepsy syndrome.

The first principle of AED use is to

select the most efficacious AED for the
seizure type or epilepsy syndrome present. With regard to AED use, seizures
are fundamentally grouped into partial
onset (including simple partial, complex
partial, and secondarily generalized),
primary generalized tonic clonic, myoclonic, atonic, and absence because
these are the categories that share efficacy for specific drugs. All partialonset seizure types can be grouped
together because AEDs useful for one
type are useful for another. It is important to also consider infantile spasms
and neonatal seizures in young children, but this chapter deals with seizures in adults and older children.
The second principle is to consider
the side effect profile of each drug and

Relationship Disclosure: Dr Fountain has received research support from NeuroPace, Inc., Medtronic, Inc.,
Johnson and Johnson, and UCB.
Unlabeled Use of Products/Investigational Use Disclosure: Dr Fountain discusses the unlabeled use of some
antiepileptic drugs as initial monotherapy. These instances are disclosed.

Copyright # 2010, American Academy of Neurology. All rights reserved.

121

" CHOOSING AMONG AEDs

KEY POINTS

122

The second
principle of
prescribing an
AED is to
consider the
side effect
profile of each
drug and
the patients
unique
characteristics,
eg, depression,
migraine, pain,
obesity, woman
of childbearing
potential,
older adult.
The third
principle of
selecting an
AED for a
patient is to
consider
convenience
and dosing.
Drugs dosed
twice per day
or less lead
to better
compliance
than drugs
dosed more
frequently.
The fourth
principle of
prescribing an
AED is to select
the lowest-cost
AED.
Traditional
drugs are least
expensive,
especially in
generic forms.
Starting and
stopping an
AED should be
done in a
specific, simple
manner.

TABLE 6-1

"

Antiepileptic
Drugs Approved
in the United
States

Conventional
Carbamazepine
Clonazepam
Clorazepate
Diazepam
Ethosuximide
Phenobarbital
Phenytoin
Primidone
Valproic acid

"

Second-Generation
Felbamate
Gabapentin
Lacosamide
Lamotrigine
Levetiracetam
Oxcarbazepine
Pregabalin
Rufinamide
Tiagabine
Topiramate
Vigabatrin
Zonisamide

the patients unique characteristics. AEDs

should be avoided if they have side
effects that would be problematic for a
specific patient. The most important
determinants are whether the patient
belongs to a special population, such
as older adults, children, or women of
childbearing potential. Important comorbid conditions that take advantage of
AED side effects or can be treated by
an AED include depression, migraine
headache, chronic pain, and obesity.
When patients fit into one of these
Continuum Lifelong Learning Neurol 2010;16(3)

categories, then selecting an AED is often straightforward because only one

or a few AEDs are appropriate.
The third principle is to consider
convenience. Drugs dosed twice per
day or less lead to better compliance
than drugs dosed more frequently. The
use of extended-release preparations of
carbamazepine, lamotrigine, and valproate allow dosing twice daily, or potentially once daily, and zonisamide and
phenobarbital can often be dosed once
per day. Some people must use or prefer liquid or sprinkle forms, so selecting
a drug available in these formulations is
logical. All conventional AEDs are available in liquid preparations as are gabapentin, levetiracetam, and oxcarbazepine.
The fourth principle is to select the
lowest-cost AED, an increasingly important consideration. Conventional drugs
are least expensive, especially in generic
forms. Phenobarbital is the least expensive at less than $20 per month and is
only dosed once per day, so it can be appropriate for indigent patients without
insurance who tolerate it. Conventional
AEDs cost $50 to $100 per month; newer
generic AEDs, $200 to $300 per month;
and brand-name second-generation
AEDs, $300 to $600 per month.
Starting and stopping an AED should
be done in a specific, simple manner.
The titration schedule recommended
for AEDs by the manufacturer is determined from clinical trials, typically
from forced titration studies. In general, AEDs are better tolerated when
started slower. Therefore, the adage
of start low and go slow is best applied. Most people find dose-escalation
schedules to be confusing. Many schedules recommend increasing the dose
every 3 days. This is problematic because 3-day increments are not logical and it is inconvenient for patients
who fill a 7-day pill box. It is much
easier for patients to remember to
change doses each week, on a specific
day of the week or on the day they fill

their pill box. It may take longer to

reach a target dose if an AED is escalated by one pill each week, but it is
more likely that it will be tolerated and
the proper dose will be achieved.
Monotherapy is recommended whenever possible because use of more AEDs
increases the risk of side effects. Drugs
should be overlapped when transitioning from one to another, but then the
first drug should be withdrawn to avoid
accumulating unnecessary drugs. The
new drug should be added and increased to the target dose to see that it
is effective and tolerated before the first
drug is withdrawn. Before a new AED is
added, the physician should be sure
the existing drug is ineffective. The only
means to determine that the drug is
ineffective is to increase the AED dose
until either seizures come under control
or side effects occur. When side effects
are induced, the dose should be reduced to the maximum well-tolerated
dose. Levetiracetam and gabapentin are
exceptions to this rule because many
patients can tolerate extremely large
doses so the maximum tolerated dose
may never be achieved. Levetiracetam is
US Food and Drug Administration (FDA)
approved up to 3000 mg/d, but some
patients get benefit from doses up to
4000 mg/d or rarely 5000 mg/d. Larger
doses are rarely useful. Gabapentin is
FDA approved to 1800 mg/d, but doses
of up to 3600 mg/d are often needed;
larger doses rarely induce side effects
but seem to be no more effective,
possibly because absorption is saturated.
If no benefit is seen from these AEDs at
these doses, it is not usually useful to
increase them even if side effects have
not occurred.
COMPARATIVE
EFFECTIVENESS STUDIES:
INITIAL MONOTHERAPY
The first and most logical question in
selecting any drug for any condition is,
Which drug is most effective? Un-

fortunately, for AEDs the answer is unknown for initial monotherapy of seizures because the question has not
been investigated sufficiently. FDAapproved indications (Table 6-2) do
not provide much guidance because
many drugs that are not approved for
monotherapy are nonetheless clearly
useful in everyday practice. A study is
needed to compare all available AEDs for
initial monotherapy in a single, very large,
randomized, blinded controlled trial,
looking separately at patients with
partial-onset seizures and those with
generalized tonic-clonic seizures. Although there is also a need for studies
in absence seizures and all other seizure types, the need is most acute for
seizures of partial-onset or unknown
type because this is the largest population of new-onset epilepsy. Unfortunately, such studies do not exist. Instead,
we must rely on: (1) individual efficacy
studies, (2) small comparative trials, and
(3) open-label comparative trials.
Individual efficacy studies of AEDs
for initial monotherapy of partial-onset
seizures have been reviewed and the
evidence categorized in an AAN practice parameter.1 These studies are useful in understanding which AEDs have
been studied for specific indications
but are not very useful in selecting
among AEDs because the data are
limited for two main reasons. (1) Some
AEDs only have class II monotherapy
studies or only have efficacy studies for
adjunctive therapy but yet are clearly
useful as monotherapy. In fact, what is
the rationale for suggesting that a drug
would be useful in the difficult circumstance of adjunctive therapy but yet
would not be useful as initial monotherapy? (2) Results from one efficacy
study cannot be directly compared to
another, even for identical study designs, because the small variability introduced in different studies is likely to
mask the small difference in efficacy
between AEDs.
Continuum Lifelong Learning Neurol 2010;16(3)

KEY POINT

Monotherapy is
recommended
whenever
possible
because use
of more AEDs
increases
the risk of
side effects.

123

" CHOOSING AMONG AEDs

TABLE 6-2

US Food and Drug Administration 2009 Approved

Indications

Antiepileptic Drug

Seizure Types

Age (y)

Monotherapy

Carbamazepine

CPS, GTC, mixed

Clonazepam

LGS, akinetic, MYO, ABS

Yes

Clorazepate

Partial

No

Diazepam

Convulsive disorders

No

Ethosuximide

ABS

Phenobarbital

NS

Phenytoin

CPS, GTC (neurosurgery)

Pregabalin

Partial onset

Adult

No

Primidone

NS

NS

Valproic acid

CPS, ABS

Yes

Multiple including ABS

No

Partial

Adults

Yes

Partial and generalized

in LGS

Children

No

Gabapentin

Partial

>3

No

Lamotrigine

Partial, generalized in
LGS, PGTC

>1

Conversion to
monotherapy
(>16 years old)

Lacosamide

Partial

>16

No

Levetiracetam

Partial, MYO, PGTC

(>6 years old)

>4

No

Oxcarbazepine

Partial

>2

Yes (>4)

Rufinamide

LGS

>3

No

Tiagabine

Partial

>12

No

Topiramate

Partial, PGTC, LGS

>2

Yes (>10)

Zonisamide

Partial

Adults

No

Vigabatrin

Infantile spasms

Children

Yes

Refractory CPS

Adults

No

Felbamate

124

CPS = complex partial seizures; GTC = generalized tonic-clonic seizures;

LGS = Lennox-Gastaut syndrome; MYO = myoclonic seizures; ABS = absence seizures;
NS = not specified; PGTC = primary generalized tonic-clonic seizures.
Adapted with permission from Fountain NB. Manual of antiepileptic drug therapy. West Islip, NY: Professional
Publications, Inc., 2010.

A few studies have compared traditional AEDs as initial monotherapy for

partial seizures. Two Department of VetContinuum Lifelong Learning Neurol 2010;16(3)

erans Affairs (VA) Cooperative Trials are

seminal studies of traditional AEDs. The
first study compared carbamazepine,

phenytoin, phenobarbital, and primidone as initial monotherapy in 622

adults.2 Drug failure was defined as
failure to control seizures after 2 years
or the occurrence of unacceptable side
effects. As in most studies that use an
intent-to-treat analysis, the drug is called
a failure if it causes unacceptable side
effects even if it controls seizures so
that efficacy is a combination of seizure control and tolerability. This study
found carbamazepine and phenytoin
to be statistically more efficacious than
primidone; phenobarbital was intermediate. The authors concluded that carbamazepine and phenytoin are drugs
of first choice for partial-onset seizures.
This study guided therapy from the
time it was published in 1985 until the
approval of second-generation AEDs beginning in 1995.
The second VA Cooperative Trial
was published in 1993 and compared
valproate to carbamazepine in 206 patients with complex partial and 237
patients with secondarily generalized
tonic-clonic seizures.3 Carbamazepine
had better efficacy than valproate for
complex partial seizures, but no difference was found for control of secondarily generalized seizures. A composite
score that combined efficacy and side
effects also favored carbamazepine. This
study supported the popular notion
that carbamazepine is preferable to valproate for complex partial seizures.
Topiramate at doses of 100 mg/d and
200 mg/d were compared to standard
drug therapy of new-onset epilepsy in
a randomized controlled trial.4 Patients
were assigned to one of two treatment
groups depending on whether under
standard clinical circumstances they
would have been given carbamazepine
(presumably for partial-onset seizures)
or valproate (presumably for generalized
seizures) No differences were found in a
variety of outcomes for the 390 patients
assigned to the group comparing carbamazepine 600 mg/d to the two topi-

ramate doses. Similarly, no difference

was found in efficacy for the 223 patients
assigned to valproate 1250 mg/d compared to the two topiramate doses. Although the overall number of subjects is
relatively large, the number assigned to
each group is relatively small given the
small difference expected between treatments. Nevertheless, it would seem that
topiramate is at least as effective as these
traditional agents.
Carbamazepine has been compared
to lamotrigine in five blinded, randomized clinical trials in new-onset epilepsy,
which were combined in a systematic
meta-analysis.5 The meta-analysis included 1384 subjects and found that
lamotrigine had the longest time to
treatment failure and carbamazepine
had better time to first seizure and rate
of seizure freedom at 6 months. In general, carbamazepine was more likely to
be withdrawn for side effects. Therefore, lamotrigine seems to be better
tolerated than carbamazepine, but among
those who tolerate carbamazepine, the
efficacy is similar to lamotrigine.
The Standard and New Antiepileptic
Drugs (SANAD) study was a two-part
study conducted in the United Kingdom to examine the outcome from the
use of common AEDs as initial monotherapy in the manner employed in
clinical practice. Therefore, it was not
blinded. One part of the study randomized 1721 patients to compare carbamazepine to four other commonly
used AEDs in patients for whom carbamazepine would be selected as the usual
standard therapy.6 Subjects were randomized, but the dose titration and final
dose was left up to investigators. Subjects who could tolerate the drugs ended
up on average daily doses of 249-mg
lamotrigine, 662-mg carbamazepine,
1496-mg gabapentin, 1019-mg oxcarbazepine, or 181-mg topiramate. Lamotrigine had the longest time to treatment
failure, and carbamazepine had the
shortest time to a 12-month remission.
Continuum Lifelong Learning Neurol 2010;16(3)

KEY POINTS

The first VA
Cooperative
Trial concluded
carbamazepine
and phenytoin
were more
efficacious
than primidone
and better
tolerated than
phenobarbital.
The second VA
Cooperative
Trial found
carbamazepine
was more
efficacious
than valproate
for complex
partial seizures.
Lamotrigine
is better
tolerated than
carbamazepine,
but among
those who
tolerate
carbamazepine,
the efficacy is
similar to
lamotrigine.

125

" CHOOSING AMONG AEDs

KEY POINTS

126

The SANAD
study found
lamotrigine and
carbamazepine
superior to
gabapentin,
oxcarbazepine,
and topiramate,
but it was an
open-label study
so the results are
questionable.
Overall, it can be
concluded that
class I evidence
is not available
to demonstrate
that any one
AED is more
efficacious
than another
for initial
monotherapy
of partial-onset
seizures.
The SANAD
study found
better efficacy
for valproate
than for
lamotrigine
or topiramate
for presumed
generalized
seizures.

The authors interpret the results to

suggest that lamotrigine is superior
to carbamazepine, as well as superior
to the other drugs tested. There does
seem to be sufficient evidence to suggest that lamotrigine and carbamazepine are superior to the other drugs.
However, the differences between lamotrigine and carbamazepine are small
for almost all measures, and there is no
difference in the drugs when analysis is
confined to the period when subjects
were on monotherapy with just the drug
to which they were randomized. This
makes it difficult to know whether the
statistical differences are clinically meaningful. The open-label nature of the
study should probably cause us to demand a greater difference in outcomes
before widely changing practice.
Overall, it can be concluded that
class I evidence is not available to
demonstrate that any one AED is clearly
more efficacious than another for initial
monotherapy of partial-onset seizures.
Carbamazepine has been the preferred
first drug, but evidence suggests that
lamotrigine is more efficacious and better tolerated than carbamazepine, gabapentin, and possibly oxcarbazepine.
Some evidence also suggests that topiramate is as efficacious as carbamazepine, but analyses have not included
levetiracetam and other newer drugs.
Thus, little evidence compels the choice
of one AED over another for treatment of
partial seizures. Other unique drug and
patient characteristics are likely to be
helpful in selecting an AED, as will be
discussed later.
Initial AED monotherapies for generalized seizures have not received the
same attention as for the treatment of
partial-onset seizures. Generalized motor seizures, principally generalized tonicclonic seizures, have been studied in
some cases, but most new-onset studies
include generalized seizures with seizures of unknown type or sometimes
with partial seizures on the theory that it
Continuum Lifelong Learning Neurol 2010;16(3)

is difficult to determine whether seizures

are partial or generalized early in the
diagnosis. The previously mentioned
AAN practice parameter has reviewed
individual efficacy studies of specific
drugs as initial monotherapy for generalized seizures.1 The parameter concludes that lamotrigine, topiramate,
and oxcarbazepine are efficacious as
initial monotherapy for a population of
mixed partial and generalized seizures
based on efficacy studies in mixed populations. Subsequently, a study demonstrated the efficacy of levetiracetam
in idiopathic generalized epilepsy, but
for adjunctive therapy rather than initial therapy.7
The second part of the SANAD study
compared valproate, lamotrigine, and
topiramate for generalized or unclassified new-onset seizures in 716 patients.8
Valproate had the greatest efficacy for
generalized seizures because of a longer
time to treatment failure than topiramate and shorter time to a 1-year
remission than lamotrigine. Thus, the
authors concluded that valproate is the
preferred drug for treatment of generalized seizures.
Absence seizures were historically
treated with ethosuximide and more
recently with valproate. In 1982, a small
double-blind comparative cross-over study
of valproate versus ethosuximide in a
mixed population of 41 patients with
new-onset or refractory absence seizures found valproate to be as effective
as ethosuximide.9 It was not powered sufficiently to find superiority of either drug,
though. The efficacy of lamotrigine for
new-onset absence seizures was demonstrated in a prospective withdrawal
protocol.10 A very recently published
prospective comparative efficacy study
compared valproate, ethosuximide, and
lamotrigine for new-onset absence seizures.11 Ethosuximide and valproate were
more efficacious than lamotrigine, and
ethosuximide had less effect on attention. This suggests that ethosuximide is

the drug of choice for new-onset absence

seizures. However, it is important to
remember that ethosuximide is not
effective against motor seizures, so valproate is usually the drug of choice when
both absence and generalized tonicclonic seizures are present. In addition,
ethosuximide, like valproate, requires
regular surveillance blood tests, while
lamotrigine does not.
COMPARATIVE EFFECTIVENESS
STUDIES: ADJUNCTIVE THERAPY
Comparative efficacy studies of adjunctive therapy for seizures suffer from
the same limitations as initial monotherapy studies. No large double-blind
placebo-controlled comparative clinical trials of all available AEDs exist.
Thus, we are again left with examining
data from individual efficacy studies,
meta-analyses, and open-label comparative studies.
Individual efficacy studies comparing the AEDs to placebo for partialonset seizures have been performed
for all second-generation and most traditional AEDs because such studies are
required for FDA approval (Table 6-2).
These are double-blind, randomized,
placebo-controlled trials as adjunctive
therapy in patients with established
epilepsy who are taking one to three
AEDs. The efficacy of adding a single
experimental drug is compared to adding placebo. Established epilepsy is
logically the first population in which
to study the efficacy of a potential AED
because this population has the greatest need and the potential benefit outweighs the risk. The usual format is to
determine whether seizure frequency
is reduced more in patients randomized to the experimental drug than to
the placebo group. Seizure frequency is
typically established in a 2-month baseline phase, and patients with at least
two (or sometimes at least four) seizures per month are randomized to
receive the study drug or placebo in a

3-month treatment phase. Seizure frequency typically decreases in a clinical

trial of AEDs, even in the placebo
group. The usual outcome measure is
to determine whether seizure frequency is reduced more in the drug-treated
group than the placebo group.
An AAN practice parameter has evaluated individual efficacy studies for
adjunctive therapy of partial seizures
similar to that reviewed for initial
monotherapy discussed above.12 Not
surprisingly, it concludes that there is
evidence of efficacy for all available
AEDs for partial-onset seizures compared to placebo. For generalized seizures, data supported only the efficacy
of adjunctive topiramate and lack of
efficacy for gabapentin. The practice
parameter does not review comparative trials or make any conclusions
about which AED is more efficacious.
Meta-analyses or comparisons of results from individual efficacy trials of
similar design have not found differences that significantly inform practice
because there are no clear winners.
A seminal careful consideration of metaanalyses of gabapentin, lamotrigine,
tiagabine, topiramate, vigabatrin, and
zonisamide found each to demonstrate superiority to placebo but no
clear demonstration that any one drug
was superior to another.13 Comparison of meta-analyses of levetiracetam,
oxcarbazepine, zonisamide, and remacemide appropriately found that remacemide alone was not likely to be
effective but did not inform practice
about drug superiority.14 A combined
method of assessing seizure control
and side effects of second-generation
AEDs reached a similar conclusion.15
Many Cochrane systematic reviews of
individual AEDs or paired comparisons have yielded similar results.
An accepted summary of the efficacy
data from adjunctive therapy of partial
and generalized seizures is that no
compelling data suggest that one AED
Continuum Lifelong Learning Neurol 2010;16(3)

KEY POINT

There is evidence
of efficacy
of all secondgeneration
AEDs for
treatment of
partial-onset
seizures, but
there is not
compelling
evidence that
one is more
efficacious than
another.

127

" CHOOSING AMONG AEDs

KEY POINTS

128

An AED may
appear more
efficacious
in an
intent-to-treat
analysis
because it
controls
seizures better
or because
it is better
tolerated.
Select an AED
based on the
presence of
depression,
migraine,
chronic pain,
obesity, woman
of childbearing
potential, or
older age.
Lamotrigine has
the best
evidence for
efficacy in
treating
depression in
patients with
epilepsy.

is better than another. If efficacy studies are not particularly useful in selecting an AED, then what is? The AAN
practice parameter and most experts
recommend considering unique patient
characteristics to guide AED selection,
such as patient age, concomitant medications, AED tolerability, safety, and
efficacy.
The discussion of efficacy for AEDs
is complicated by the common occurrence of significant side effects, which
are especially common and cumbersome for the conventional AEDs. Comparative efficacy studies take this into
account by use of an intent-to-treat
analysis. This method logically considers a subject a treatment failure if
the drug is inadequate to control seizures. The straightforward situation is
when subjects persist in having seizures despite reaching the target dose
of the drug. However, subjects would
also be considered a treatment failure
if they could not tolerate the drug
because of side effects. Thus, the efficacy measure is always a combination
of the drugs ability to control seizures
and its tolerability. A drug may appear
better because it actually controls seizures better or because it has more
tolerable side effects.
It is clear that some drugs more
commonly induce side effects than
others. For example, in the second VA
Cooperative Study, weight gain, hair
loss, and tremor were more common
in the valproate-treated group, while
rash was much more common with
carbamazepine. Other studies have found
that carbamazepine-treated patients have
more side effects than lamotriginetreated patients.16,17 This might lead
to the conclusion that lamotrigine is
preferable to carbamazepine. However,
the intent-to-treat analysis compensates
for this and only when a drug wins in
the analysis is it actually better for a
whole population, regardless of whether
it is better tolerated. The situation is
Continuum Lifelong Learning Neurol 2010;16(3)

different when selecting drugs for individual patients where the unique characteristics of the patient and the drug
can be taken into account. Each AED
has common side effects that will not be
discussed in detail here except in relation to AED selection in the following
section.
UNIQUE PATIENT AND
ANTIEPILEPTIC DRUG
CHARACTERISTICS
Unique patient characteristics useful
in selecting an AED include the presence of comorbidities or being a member of a special population (Table 6-3)
(Case 6-1). Patients with epilepsy have
many comorbidities, such as depression, migraine, chronic pain, and obesity, which can be treated with an AED
for dual purposes. The presence or predisposition to some comorbidities suggests that some AEDs should be avoided
because their side effects might be
particularly problematic for a particular
patient; for example, the presence of
a history of kidney stones suggests
topiramate and zonisamide should be
avoided. Similarly, if a patient is a member of a special population, such as
women of childbearing potential or
older adults, AEDs can be selected that
are known to be useful in this group
or AEDs can be avoided when they
cause problems unique to these groups.
Depression is very common in epilepsy, especially temporal lobe epilepsy.
Lamotrigine has demonstrated beneficial effects for patients with epilepsy.
Lamotrigine reduces depression in prospective randomized trials when added
to existing AED therapy and reduces
anger, hostility, and depression better
than levetiracetam.18,19 Valproate has
been historically associated with treating comorbid psychiatric disease because of its use in acute mania. Evidence suggests, however, that lamotrigine
more effectively improves depression
than valproate does.20 Nevertheless, if

TABLE 6-3

Antiepileptic Drug Preferences in Special Circumstances

Patient Characteristics

Antiepileptic Drug Preferences

Depression

Lamotrigine, oxcarbazepine

Migraine

Topiramate, valproate

Chronic pain

Pregabalin, gabapentin, oxcarbazepine,

carbamazepine, lacosamide

Obesity

Topiramate, zonisamide
Avoid pregabalin, gabapentin, valproate

Woman of childbearing
potential

Avoid valproate

Older adult

Lamotrigine, gabapentin, topiramate

Asian

Avoid carbamazepine

Nephrolithiasis

Avoid topiramate and zonisamide

Atopic (rash prone)

Avoid lamotrigine, carbamazepine

coincidental bipolar disorder is present, then lamotrigine and valproate are

good choices. Oxcarbazepine reduced
depression in an uncontrolled study
and is also useful in bipolar disorder,
similar to carbamazepine, so it is also a
reasonable consideration in this circumstance.21
Levetiracetam can cause irritability and
anger with psychiatric adverse events

occurring in 10% of patients in a retrospective review.22 However, the actual

rate of levetiracetam-induced psychiatric side effects is not known because
placebo-controlled clinical trials did
not specifically investigate these types
of side effects, so they may have
gone unnoticed. In studies used for
FDA registration, irritability did not
emerge as a very common event in

Case 6-1
A 30-year-old male construction worker presented with new-onset complex
partial seizures and focal left temporal sharp waves and normal MRI.
Which AED should be prescribed?
Comment. Any conventional AED could be selected, except
ethosuximide, which is useful only for absence seizures, or almost any
second-generation AED. This patient has few unique characteristics that
suggest one drug is more appropriate than another, based on the
information available. It would be important to ask specifically about
comorbid conditions that would influence drug choice. If he has migraine
headaches, then topiramate would be appropriate unless he has a
history of kidney stones. If he has depression, then consider lamotrigine.
If mood instability is an issue, then consider oxcarbazepine. If cost is an
issue, then consider a conventional AED, especially carbamazepine; or
consider phenobarbital if compliance is likely to be an issue. Levetiracetam
is not FDA approved for initial monotherapy but is often prescribed
because it is easy to use, well tolerated, and has a wide therapeutic index.

Continuum Lifelong Learning Neurol 2010;16(3)

129

" CHOOSING AMONG AEDs

KEY POINT

130

Topiramate and
valproate have
the best
evidence for
efficacy against
migraine
headache in
patients with
epilepsy.

levetiracetam-treated patients, but this

type of symptom is unlikely to be captured during routine efficacy trials and
is probably not a very high-frequency
event. Although the overall rate of
irritability is seemingly low, it can be
significant when it occurs. Therefore, it
seems prudent to avoid levetiracetam in
patients for whom irritability would be
particularly problematic. Topiramate has
been reported to cause depression, possibly dependent on the dose-escalation
rate, but is also being investigated for
the treatment of resistant depression,
so its role in treating comorbid depression is minimal.23
Migraine is common in patients with
epilepsy. Valproate has been the historical mainstay for migraine prophylaxis,
including in patients with epilepsy. However, valproate is teratogenic and carries the risk of hepatitis and other side
effects that necessitate routine blood
work. Topiramate also has efficacy in
migraine prophylaxis, but because it
lacks the end-organ toxicity and apparently the teratogenicity of valproate,
it is often preferred over valproate
when no other mitigating circumstances are present. Open-label or singleblind studies have suggested improved
headache control with zonisamide,24
pregabalin,25 and gabapentin,26 as has a
case series of levetiracetam,27 but
double-blind placebo-controlled studies for these drugs are lacking. Oxcarbazepine was not effective for migraine
prophylaxis in a placebo-controlled
trial.28 Lamotrigine was less efficacious
for migraine than topiramate in a small
double-blind placebo-controlled trial
and only marginally better than placebo,29 confirming the general clinical
impression that lamotrigine has low efficacy for migraine.
Chronic and neuropathic pain is
not particularly common in patients
with epilepsy, but it is relatively common in the general population, so
epilepsy patients with comorbid pain
Continuum Lifelong Learning Neurol 2010;16(3)

syndromes are seen somewhat frequently. Some AEDs have clearly demonstrated utility for pain, while others
lack data. Pain syndromes found to
be responsive to AEDs include diabetic
neuropathy, postherpetic neuralgia, trigeminal neuralgia, spinal cord injury
pain, phantom limb pain, fibromyalgia,
and cancer pain. Somewhat consistent
efficacy across pain syndromes has been
found for carbamazepine, oxcarbazepine, gabapentin, and pregabalin, with
little effect of topiramate and inconsistent effect of lamotrigine, as outlined in
an evidence-based review.30 Gabapentin
and pregabalin have been particularly
well studied, especially for postherpetic
neuralgia, and pregabalin for fibromyalgia. Peripheral neuropathy pain has
traditionally been treated with carbamazepine, but randomized controlled
trials demonstrate efficacy of gabapentin,
pregabalin, and lacosamide, although
not all of these drugs have regulatory
approval for this indication. Topiramate
has not generally been efficacious for
pain syndromes.
Obesity is a common problem, especially in people with epilepsy who
are less active. Pregabalin, gabapentin,
and valproate can cause weight gain
and should be avoided when that possibility is a problem. Topiramate and
zonisamide can cause weight loss, so
this characteristic may be useful for
obese patients. Conversely, it means
that these drugs should be avoided in
cachectic patients.
Women of childbearing potential
have several considerations in selecting
an AED. The most important issue is
the avoidance of highly teratogenic
AEDs. Pregnancy registries have demonstrated that valproate is associated
with a major congenital malformation
rate of approximately 10%, especially at
high doses.31 The risk of birth defects
in the general population is approximately 2%. Lamotrigine is probably not
associated with an increase in the overall

risk of birth defects, but among the birth

defects that develop, cleft lip and palate are more common. Surprisingly
little systematic data have been collected on older AEDS,32 but it appears
carbamazepine is not associated with a
significant increase in teratogenicity.
The risk for other AEDs is not well
known because registries have not yet
accumulated enough pregnancies. However, the risk does not appear to be high
for topiramate and levetiracetam. The
rates for phenobarbital and phenytoin
are still relatively unknown using contemporary study methods. Overall, it is
reasonable to avoid valproate in women of childbearing potential and consider using lamotrigine, which has a
known, but low, rate of malformation
(Case 6-2).
Elderly patients constitute a special
population because they are more
likely to be on multiple medications,
have other systemic illnesses, and tolerate drugs less well than younger patients with epilepsy. Carbamazepine
has been the traditional drug of choice,
but like phenytoin and phenobarbital,
it induces hepatic enzymes, which is
a particular problem in older adults
because of potential drug interactions.
In addition, these drugs have a narrow
therapeutic index. Some randomized
controlled trials in older adults have

not found lamotrigine to have greater

efficacy than carbamazepine or gabapentin, but lamotrigine is better tolerated, leading some to conclude that it
is preferable.16,33 A placebo-controlled
study of low-dose topiramate (50 mg)
and high-dose topiramate (200 mg) in
older adults found them to be equally
efficacious as initial monotherapy, but
the higher dose was more efficacious
as adjunctive therapy.34 These data indicate that topiramate is better than
placebo but do not indicate whether it
is better than lamotrigine or alternative
drugs. Other AEDs have not had controlled comparative trials in older adults.
Thus, lamotrigine, gabapentin, topiramate, and carbamazepine are efficacious
in older adults and reasonable considerations. Which of these drugs is better
remains unknown, but the secondgeneration agents, especially lamotrigine
and topiramate, are better tolerated
(Case 6-3).
A few drugs have idiosyncratic side
effects that occasionally influence drug
choice. Topiramate and zonisamide are
associated with kidney stones. Although
it is not known whether a history of
kidney stones predisposes to developing them again in patients treated with
these drugs, it is prudent to avoid
these drugs in patients with a history
of kidney stones.

KEY POINTS

Valproate,
gabapentin,
and pregabalin
can cause
weight gain,
and topiramate
and zonisamide
can cause
weight loss.
Conventional
enzyme-inducing
AEDs, such as
phenytoin and
carbamazepine,
should be
avoided in
older adults.

131
Case 6-2
A 21-year-old woman presented with the recent onset of generalized
tonic-clonic seizures after a week of sleep deprivation during spring
break. She also had occasional early morning arm jerks, and her EEG
demonstrated generalized polyspike-and-wave discharges, suggesting that
she has juvenile myoclonic epilepsy.
Comment. Valproate should be avoided because she is likely to need
long-term AED use and valproate is significantly teratogenic. Lamotrigine
has an established low risk of teratogenicity and is often selected.
However, lamotrigine has a very long dose-escalation phase to avoid rash.
Topiramate is useful although the low teratogenic risk is not as well
established. Levetiracetam is efficacious but has only been studied as
adjunctive therapy, and the teratogenic risk is also not well established.

Continuum Lifelong Learning Neurol 2010;16(3)

" CHOOSING AMONG AEDs

Case 6-3
A 67-year-old man presents with new-onset complex partial seizures due
to cryptogenic epilepsy. Which AED should be prescribed?
Comment. Lamotrigine is well tolerated and likely to have few drug
interactions. Topiramate, oxcarbazepine, pregabalin, and levetiracetam
are also often selected for similar reasons although levetiracetam has not
been studied as initial monotherapy. Carbamazepine is a consideration
if cost is an issue, but the patient will also incur the cost of blood work
that would be avoided with the second-generation medications noted
previously. Phenytoin should be avoided because of potential for drug
interactions and it is poorly tolerated in older adults.

132

Rashes in patients with epilepsy are

usually mild and, indeed, are probably
most often unrelated to an AED. These
drugs, however, can cause serious rash,
such as Steven-Johnson syndrome or
toxic epidermal necrolysis. Lamotrigine
was commonly associated with rash in
clinical trials (1:50 children), but the
incidence has decreased dramatically
with the practice of starting at low
doses and using slow dose escalation.
No reliable risk factors have been
identified that will predict who will
develop rash, but it is prudent to avoid
lamotrigine in atopic patients. If such
patients are treated with lamotrigine
and do develop a rash, it will be difficult to determine whether lamotrigine is responsible. Carbamazepine
is associated with rash, particularly
in patients with a human leukocyte
antigen-B1502 genotype. This genotype is particularly common in the
Asian population. Therefore, it has been
recommended to check this laboratory
test prior to starting carbamazepine in
Asian patients or to avoid carbamazepine in this group.
A few AEDs should not be selected
as initial monotherapy and are probably best prescribed by epilepsy specialists because of potential problems.
Vigabatrin is associated with progressive visual field loss. Felbamate was
associated with hepatitis and aplastic
anemia within 1 year of its approval,
although no additional cases have been
Continuum Lifelong Learning Neurol 2010;16(3)

reported since then. Rufinamide is not

known to be associated with any particular problems and is approved for
Lennox-Gastaut syndrome. It has efficacy for refractory partial seizures
but is best reserved for treatment of
refractory epilepsy until enough patients have been treated to determine
whether rare significant side effects will
occur.
CONCLUSION
The few comparative efficacy studies
of AEDs for the treatment of partial onset seizures do not convincingly
demonstrate that one drug is clearly
more efficacious than others for initial monotherapy in this seizure type.
Although lamotrigine has received the
most attention and appears to be
better tolerated than carbamazepine,
other drugs have not been examined
carefully. Thus, an AED should be selected on the basis of comorbid conditions or patient characteristics. Similar
issues exist for generalized seizures.
Although valproate, topiramate, and
lamotrigine have been most widely
compared, the presence of comorbid
conditions and special circumstances
are likely to be more useful in guiding selection of an AED. Taking time
to learn the principles outlined here
will considerably simplify the selection
of an AED and allow the neurologist
to approach selection of an AED
systematically.

REFERENCES
1. French JA, Kanner AM, Bautista J, et al. Efficacy and tolerability of the new
antiepileptic drugs I: treatment of new onset epilepsy: report of the Therapeutics and
Technology Assessment Subcommittee and Quality Standards Subcommittee of the
American Academy of Neurology and the American Epilepsy Society. Neurology
2004;62(8):12521260.
2. Mattson RH, Cramer JA, Collins JF, et al. Comparison of carbamazepine,
phenobarbital, phenytoin, and primidone in partial and secondarily generalized
tonic-clonic seizures. N Engl J Med 1985:313(3):145151.
3. Mattson RH, Cramer JA, Collins JF; The Department of Veterans Affairs Epilepsy
Cooperative Study No. 264 Group. A comparison of valproate with carbamazepine for
the treatment of complex partial seizures and secondarily generalized tonic-clonic
seizures in adults. N Engl J Med 1992;327(11):765771.
4. Privitera MD, Brodie MJ, Mattson RH, et al. Topiramate, carbamazepine and valproate
monotherapy: double-blind comparison in newly diagnosed epilepsy. Acta Neurol
Scand 2003;107(3):165175.
5. Gamble C, Williamson PR, Chadwick DW, Marson AG. A meta-analysis of individual
patient responses to lamotrigine or carbamazepine monotherapy. Neurology
2006;66(9):13101317.
6. Marson AG, Al-Kharusi AM, Alwaidh M, et al. The SANAD study of effectiveness
of carbamazepine, gabapentin, lamotrigine, oxcarbazepine, or topiramate for
treatment of partial epilepsy: an unblinded randomised controlled trial. Lancet
2007;369(9566):10001015.
7. Berkovic SF, Knowlton RC, Leroy RF, et al; Levetiracetam N01057 Study Group.
Placebo-controlled study of levetiracetam in idiopathic generalized epilepsy.
Neurology 2007;69(18):17511760.
8. Marson AG, Al-Kharusi AM, Alwaidh M, et al. The SANAD study of effectiveness
of valproate, lamotrigine, or topiramate for generalised and unclassifiable
epilepsy: an unblinded randomised controlled trial. Lancet 2007;369(9566):
10161026.
9. Sato S, White BG, Penry JK, et al. Valproic acid versus ethosuximide in the treatment of
absence seizures. Neurology 1982;32(2):157163.
10. Frank LM, Enlow T, Holmes GL, et al. Lamictal (lamotrigine) monotherapy for typical
absence seizures in children. Epilepsia 1999;40(7):973979.
11. Glauser TA, Cnaan A, Shinnar S, et al; Childhood Absence Epilepsy Group.
Ethosuximide, valproic acid, and lamotrigine in childhood absence epilepsy.
N Engl J Med 2010;362(a):790799.
12. French JA, Kanner AM, Bautista J, et al. Efficacy and tolerability of the new
antiepileptic drugs II: treatment of refractory epilepsy: report of the Therapeutics
and Technology Assessment Subcommittee and Quality Standards Subcommittee
of the American Academy of Neurology and the American Epilepsy Society.
Neurology 2004;62(8):12611273.

Continuum Lifelong Learning Neurol 2010;16(3)

133

" CHOOSING AMONG AEDs

13. Marson AG, Kadir ZA, Hutton JL, Chadwick DW. The new antiepileptic drugs:
a systematic review of their efficacy and tolerability. Epilepsia 1997;38(8):
859880.
14. Marson AG, Hutton JL, Leach JP, et al. Levetiracetam, oxcarbazepine, remacemide and
zonisamide for drug resistant localization-related epilepsy: a systematic review.
Epilepsy Res 2001;46(3):259270.
15. Cramer JA, Ben Menachem E, French J. Review of treatment options for refractory
epilepsy: new medications and vagal nerve stimulation. Epilepsy Res 2001;47(12):
1725.
16. Rowan AJ, Ramsay RE, Collins JF, et al. New onset geriatric epilepsy: a randomized
study of gabapentin, lamotrigine, and carbamazepine. Neurology 2005;64(11):
18681873.
17. Brodie MJ, Richens A, Yuen AW. Double-blind comparison of lamotrigine and
carbamazepine in newly diagnosed epilepsy. UK Lamotrigine/Carbamazepine
Monotherapy Trial Group. Lancet 1995;345(8948):476479.
18. Ettinger AB, Kustra RP, Hammer AE. Effect of lamotrigine on depressive symptoms
in adult patients with epilepsy. Epilepsy Behav 2007;10(1):148154.
19. Labiner DM, Ettinger AB, Fakhoury TA, et al. Effects of lamotrigine compared
with levetiracetam on anger, hostility, and total mood in patients with partial
epilepsy. Epilepsia 2009;50(3):434442.
20. Edwards KR, Sackellares JC, Vuong A, et al. Lamotrigine monotherapy improves
depressive symptoms in epilepsy: a double-blind comparison with valproate. Epilepsy
Behav 2001;2(1):2836.
21. Mazza M, Della Marca G, Di Nicola M, et al. Oxcarbazepine improves mood in patients
with epilepsy. Epilepsy Behav 2007;10(3):397401.
22. Mula M, Trimble MR, Yuen A, et al. Psychiatric adverse events during levetiracetam
therapy. Neurology 2003;61(5):704706.
23. Mula M, Hesdorffer DC, Trimble M, Sander JW. The role of titration schedule of
topiramate for the development of depression in patients with epilepsy. Epilepsia
2009;50(5):10721076.

134

24. Bermejo PE, Dorado R. Zonisamide for migraine prophylaxis in patients refractory to
topiramate. Clin Neuropharmacol 2009;32(2):103106.
25. Calandre EP, Garcia-Leiva JM, Rico-Villademoros F, et al. Pregabalin in the treatment
of chronic migraine: an open-label study. Clin Neuropharmacol 2009;[Epub ahead
of print].
26. Mathew NT, Rapoport A, Saper J, et al. Efficacy of gabapentin in migraine
prophylaxis. Headache 2001;41(2):119128.
27. Cochran JW. Levetiracetam as migraine prophylaxis. Clin J Pain. 2004;20(3):
198199.
28. Silberstein S, Saper J, Berenson F, et al. Oxcarbazepine in migraine headache: a
double-blind, randomized, placebo-controlled study. Neurology 2008;70(7):
548555.

Continuum Lifelong Learning Neurol 2010;16(3)

29. Gupta P, Singh S, Goyal V, et al. Low-dose topiramate versus lamotrigine in migraine
prophylaxis (the Lotolamp study). Headache 2007;47(3):402412.
30. Goodyear-Smith F, Halliwell J. Anticonvulsants for neuropathic pain: gaps in the
evidence. Clin J Pain 2009;25(6):528536.
31. Wyszynski DF, Nambisan M, Surve T, et al. Increased rate of major malformations
in offspring exposed to valproate during pregnancy. Neurology 2005;64(6):
961965.
32. Fountain NB. A pregnant pause to consider the teratogenicity of topiramate. Epilepsy
Curr 2009;9(2):3638.
33. Brodie MJ, Overstall PW, Giorgi L; The UK Lamotrigine Elderly Study Group.
Multicentre, double-blind, randomized comparison between lamotrigine and
carbamazepine in elderly patients with newly diagnosed epilepsy. Epilepsy Res
1999;37(1):8187.
34. Ramsay RE, Uthman B, Pryor FM, et al. Topiramate in older patients with
partial-onset seizures: a pilot double-blind, dose-comparison study. Epilepsia
2008;49(7):11801185.

135

Continuum Lifelong Learning Neurol 2010;16(3)