Seizure (2006) 15, 165—176

www.elsevier.com/locate/yseiz

Are there potential problems with generic

substitution of antiepileptic drugs?
A review of issues
P. Crawford a,*, M. Feely b,1, A. Guberman c,2, G. Kramer d,3

a
York Hospital, Wigginton Road, York YO31 8HE, UK
b
Leeds General Infirmary, Great George Street, Leeds LS1 3EX, UK
c
Ottawa Hospital, 501 Smyth Road, Ottawa, Canada K1H 8L6
d
Swiss Epilepsy Center, Bleulerstr. 60, Zurich CH-8008, Switzerland

Received 31 January 2005; received in revised form 6 December 2005; accepted 28 December 2005

KEYWORDS Summary In response to increasing cost pressures, healthcare systems are encoura-
Antiepileptic drugs; ging the use of generic medicines. This review explores potential problems with
Generic substitution; generic substitution of antiepileptic drugs (AEDs).
Bioequivalence; A broad search strategy identified approximately 70 relevant articles. Potential
Pharmacokinetics; problems with generic substitution included:
Review
 potentially serious consequences of failure of therapy, particularly in well-controlled
patients;
 potential for adverse events and variability of response to AEDs;
 need for careful titration and dosing of AEDs and susceptibility of some patients to
develop problems, even with small changes in drug levels;
 bioequivalence, as defined by regulatory bodies, may not correspond to therapeutic
equivalence for AEDs, because of the permitted range of bioavailability for generics,
evaluation methods that use small numbers of relatively young healthy volunteers
and individual variation;
 potential for problems from poor continuity of supply;
 cost savings may be outweighed by the cost of adverse consequences;
 potential medico-legal consequences in patients who did not give informed consent
to switching of AEDs.

* Corresponding author. Tel.: +44 1904 725753/4/5; fax: +44 1904 726374.
E-mail addresses: pamela.crawford@york.nhs.uk (P. Crawford), feelym@leedsth.nhs.uk (M. Feely),
aguberman@ottawahospital.on.ca, aguberman@rogers.com (A. Guberman), g.kraemer@swissepi.ch (G. Kramer).
1
Tel.: +44 113 3922702/3472/3470; fax: +44 113 2423811.
2
Tel.: +1 613 737 8142; fax: +1 613 737 8857.
3
Tel.: +41 1 387 6302; fax: +41 1 387 6396.

1059-1311/$ — see front matter # 2006 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.seizure.2005.12.010
166 P. Crawford et al.

The limited evidence (mainly case reports with some pharmacokinetic studies)
appears to support these concerns for older AEDs. As a result, restrictions on use of
specific generic AEDs are in place in some countries and recommended by some lay
epilepsy organisations.
As more AEDs lose patent protection, it is important to examine the question of
whether generic substitution may pose problems for patients with epilepsy, and
whether there should be safeguards to ensure that both physician and patient are
informed when generic substitution occurs.
# 2006 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.

Introduction As a consequence, the proportion of prescriptions

Healthcare systems in the developed world are
under pressure from increasing numbers of elderly
people, increasing patient expectations, and costly
advances in healthcare technology including new
medicines. Without the economic growth to meet
these increasing demands, most healthcare systems
have taken steps to limit cost escalation, including
the introduction of health technology assessment
agencies to determine the cost-effectiveness of new
therapies, pricing controls such as referencing to
the cost of treatments in other countries, and
encouraging the use of cheaper generic medicines
instead of branded products. While the economic
need to limit healthcare costs is not questioned, it is
important to ensure that patient health is not com-
promised. This review aims to explore potential
problems with generic substitution of antiepileptic
drugs (AEDs).
Generic prescribing
The innovations from research-based pharmaceuti-
cal companies are protected from copying by
patents that typically last 20 years from the first
filing of the new chemical entity. As all the devel-
opment work and clinical trials on a new chemical
entity is conducted by the innovator in order to gain
the initial marketing authorisation, the generic
manufacturer has only to demonstrate bioequiva-
lence of the generic product and the branded pro-
duct to gain a product licence. This is the main
reason why generic products can be priced well
below the price of innovative products.1
Since generics are the same chemical entity as
the branded product, healthcare payers and many
physicians expect that generic products can be used
interchangeably with each other and the original
product.2 Consequently, generic prescribing is
widely advocated as a measure of cost containment.
Physicians are routinely encouraged by healthcare
management to prescribe products generically,3
while computer prescribing systems often automa-
tically convert brand names into generic versions.
written generically has increased markedly over
recent years.4
In some markets (e.g. UK), generic products can
only be dispensed against prescriptions that are
written generically–—substitution of a generic pro-
duct for a prescription written as the brand name is
not permitted. However, in many other markets
(e.g. Canada), generic substitution (i.e. dispensing
of a generic version if one is available, even if the
physician used the brand name on the prescription)
is allowed or mandated. If the physicians do not
want the generic product to be given, they may be
able to indicate this on the prescription or by filling
in other exemption forms. In some countries, if
patients insist on receiving the branded product,
they are expected to pay the difference in cost over
the lowest price generic product.
Although healthcare systems tend to favour the
use of generic products because of the savings in
cost of medicines, there are a number of potential
disadvantages to the use of generic products
(Table 1).3,5 In the majority of therapeutic areas,
these disadvantages are irrelevant or minor in com-
parison to the economic benefits. However, many
Table 1 Disadvantages of using a generic product3,5
Rate and extent of absorption (bioavailability) may
differ between generics and branded products
Generic names are not as easy to remember,
spell or pronounce as branded names
Generic products usually differ in appearance
(e.g. colour, shape) from the brand and from one
another, causing confusion and anxiety for patients
Excipients and colorants used in generic products
may differ from the brand–—although these agents
are intended to be inert, they can cause problems in
some patients
If problems occur with a generic product, it may
be difficult to identify the manufacturer or supplier,
once it has been dispensed, and the innovator
company may be the recipient of the
pharmacovigilance report rather than the generic
company
Use of generic products may inhibit research and
innovation
Are there potential problems with generic substitution of antiepileptic drugs
healthcare systems recognise that changing the
supplier of medicines could compromise care of
patients with certain conditions.6 Physicians are
advised to prescribe these products by brand name,
and such products are exempt from generic substi-
tution. A number of experts and professional bodies
have recommended caution with generic substitu-
tion of drugs required for epilepsy.7—9
Search strategy
In order to review the issue of generic substitution of
antiepileptic drugs, a systematic literature search
was conducted. A broad search strategy was
adopted that included the generic names of AEDs,
together with terms relevant to generic substitution
such as bioequivalence, bioinequivalence, bioavail-
ability, pharmacokinetics and narrow therapeutic
index. Desktop databases were searched, including
Medline, Current Contents and International Phar-
maceutical Abstracts. Approximately 1470 refer-
ences were identified and approximately 70
relevant articles were selected as these were the
only ones which related to epilepsy and generic
prescribing of antiepileptic drugs.
It was of interest to note that very few articles
described randomised controlled clinical trials com-
paring generic and branded products. The majority
of articles consisted of case reports, letters discuss-
ing case reports, or opinion pieces without presen-
tation of new data. The paucity of evidence means
that this review can only highlight the issues and
need for more carefully conducted studies.
Issues of concern
Reviewing the literature, issues of concern about
the use of generic AEDs centred around a number
of themes (Table 2), each of which is discussed in
more detail.
Characteristics of epilepsy
Epilepsy is a chronic disorder that often requires
lifelong treatment. Avoidance of seizures is the
primary goal, while keeping adverse effects to a
minimum.10 When long-term remission has been
achieved, it becomes important to avoid even a
single breakthrough seizure. Many papers high-
lighted the importance of this issue.5,7,9,11—19 Just
one seizure after a period of control can have major
implications at the social level (e.g. loss of driving
licence, loss of employment) and at the personal
level (e.g. risk of injury, loss of self-esteem). There
167
Table 2 Issues of concern about generic substitution
of AEDs
Characteristics of epilepsy–—seriousness of therapy
failure
Characteristics of AEDs–—potential for adverse events,
narrow therapeutic index, individual variation in
response
Complexity of management regimens–—need for slow
titration, drug interactions
Bioequivalence vs. therapeutic equivalence–—
permitted range of bioavailability, evaluation of
bioequivalence, individual variation
Continuity of supply–—initial prescribing vs. switching
to a generic, changes in suppliers over time
Economic value–—potential savings versus potential
costs
Legal situation and informed consent–—implications if
generic substitution without informed consent
results in seizure/adverse events
may even be fatal consequences–—the risk of death
in patients with uncontrolled seizures is higher than
in seizure-free patients.20 There is, therefore, con-
siderably more at stake when treating epilepsy than
with many other conditions, such as peptic ulcer,
when any slight loss of efficacy on changing product
would be expected to have limited consequences.
Characteristics of antiepileptic drugs
There is a limited armamentarium of treatments
for epilepsy, with drugs such as carbamazepine,
sodium valproate and phenytoin forming the main-
stay of therapy. The narrow therapeutic index of
these AEDs necessitates some monitoring of serum
drug concentrations.
Many AEDs are considered to be treatments
with a narrow therapeutic index (i.e. only a small
relative difference in dose between therapeutic
and toxic effects). A narrow therapeutic index
implies that slight variations in drug absorption
could result in significant negative health out-
comes.21 The FDA considers a drug to have a narrow
therapeutic index if:
 there is less than a two-fold difference between
the minimum toxic concentration and the mini-
mum effective concentration;
 safe and effective use requires careful titration
and patient monitoring.8,22
Narrow therapeutic index has also been used to
describe medications that practitioners consider
may present difficulties with generic substitution.8
Publications concerning AEDs have focused on
168 P. Crawford et al.

Table 3 Pharmacokinetic characteristics of AEDs that may present problems during generic substitution23
AED Factors increasing likelihood of problems with generic substitution
Low water Narrow therapeutic Nonlinear
solubility range pharmacokinetics
Phenytoin Yes Yes Yes
Carbamazepine Yes Yes No
Valproate No Yes Yes

AEDs that have a narrow therapeutic index Bioequivalence versus therapeutic

(Table 3).8,14,15,23
With all AEDs, there is wide variation in thera-
peutic dose among individuals. To ensure that the
optimal dose of AED is used and to avoid toxicity,
dosage is usually adjusted over a considerable per-
iod of time, according to the response of the indi-
vidual.10 Even AEDs with low toxicity and with a
wide therapeutic index, such as lamotrigine, require
titration to the optimal dose.24
Complexity of management regimens
Establishing seizure control can be difficult, and a
number of AEDs may have to be tried at various
doses to identify which treatment is most effec-
tive and tolerable. If adequate control of seizures
is not achieved with a tolerable dose of one or two
AEDs in sequential monotherapy, then other AEDs
are often added as adjunctive therapy, again with
titration according to the therapeutic response. As
a result, many patients with epilepsy are on a daily
regimen of multiple treatments that has been
carefully adjusted to obtain the optimal response.
Many AEDs induce hepatic microsomal enzymes,
increasing the rate of metabolism of concomitant
drugs. This further complicates the dosage of AEDs
in patients receiving polypharmacy, adding to con-
cerns about the potential for adverse conse-
quences if any aspect of this finally balanced
therapy is inadvertently disturbed. Consequently,
there is considerable reluctance from physicians
and patients to change therapeutic agents in some
patients, particularly once stability has been
achieved.5,12,23
equivalence
There are strict regulations covering the licensing of
generic products, in order to give physicians and
patients confidence that they are equivalent to
branded products. In most countries, generic pro-
ducts must have the same dose and form as the
brand. In USA, Canada and other developed coun-
tries, generic products also have to demonstrate
equivalent bioavailability (similar blood concentra-
tion profile over time) to the brand in order to obtain
a product licence (Table 4). The excipients and
manufacturing processes for generic products can
differ from the innovator product. This is why there
may be differences in the appearance, taste and
shelf life of generic and branded products.24
Furthermore, the salt or ester of the active ingre-
dient may also sometimes differ between branded
and generic products.25
In the majority of therapeutic areas, differences
within the permitted range have negligible effects,
so that equivalent bioavailability indicates bioequi-
valence (i.e. a similar effect) in the patient.5
Indeed, bioavailability is considered by licensing
authorities to be a demonstration of bioequiva-
lence. However, commentators have noted that
bioequivalence implies but does not guarantee that
a drug will have the same therapeutic and adverse
effects as the reference drug.5,24 Furthermore,
despite the lack of specific regulations concerning
the excipients in a generic formulation, these sub-
stances cannot be considered inactive or inert mole-
cules, as different salts of the same active drug can
have distinct chemical and biological properties.1

Table 4 Regulations covering licensing of generic products1,6,24

Based on single dose administration of generic and innovator brand in crossover design study carried out in
healthy normal adults (number required differs between countries: 12 in Canada; 24—36 in the USA;
18—24 recommended by the World Health Organisation)
Pharmacokinetic parameters of generic and innovator brand measured:
Area under the concentration curve over time (AUC), to indicate the amount of drug absorbed
Peak concentration (Cmax) to indicate the rate of absorption
90% confidence intervals for the geometric mean of the generic product for each of these parameters
must be within 80—125% of the corresponding parameters for the innovator brand
Are there potential problems with generic substitution of antiepileptic drugs
This potential for differences in therapeutic respo-
nse to AEDs, even though products are defined as
bioequivalent, is a recurring theme in the litera-
ture,1,2,5,6,19,23,24,26—31 with a number of issues
highlighted.
Permitted range of bioavailability
If the bioavailability of all approved products were
plotted, the branded product would be in the middle
of the distribution.8 However, once a number of
generic products are on the market, patients may
be switched from one generic to another and pos-
sibly from one that has bioavailability at the top of
the acceptable range to one that is at the opposite
end of the range, increasing the potential for clin-
ical problems. Many clinicians have raised concerns
about the range of bioavailability that is permitted
within the definition of bioequivalence.2,7,12,23,32,33
Evaluation of bioequivalence
Bioequivalence studies are usually carried out with
single doses on small numbers of healthy volunteers
(usually young adults) who are not receiving other
therapies. This is done in order to eliminate factors
(such as the presence of a disease state) that may
cause variations in the results. In the clinical situa-
tion, patients with epilepsy often receive multiple
drugs. Many patients are likely to be older or
younger than the adults used in standard tests, with
consequent differences in drug handling character-
istics. Some commentators have raised concerns
about the accepted methods for evaluating bioe-
quivalence.2,5,32,33
Individual variation
Another important concern emerging in the litera-
ture is the considerable variation in response among
individuals.2,9,19,26,34 The evaluation of bioequiva-
lence in a population of patients with epilepsy is not
the same as establishing bioequivalence for an
individual patient with epilepsy. It is well known
that patients with epilepsy respond to AEDs in
different ways, so that individual titration of doses
is required. As with other concerns, there is little
specific concrete supportive evidence in the litera-
ture. However, the American Academy of Neurol-
ogy9 noted that the ratio of generic to branded
bioavailability in individual subjects reported to
the FDA varied from 74% to 142%.
Continuity of supply
Another theme that emerges from the literature is
concern over continuity of supply.2,8,14 In general,
commentators are not against generic AEDs per se.
Rather the concern expressed is over the potential
169
consequences of having multiple suppliers of a pro-
duct. Although patients may receive a generic AED
as the first treatment following diagnosis, more
commonly, patients are switched to a generic pro-
duct during maintenance therapy.8
Regulatory authorities have recognised that the
issue of bioequivalence is different for these two
patient populations. In the USA, the FDA has used
the term prescribability to describe the use of a
generic as first therapy, and the term switchability
for the use of a generic in a patient already
established on therapy.8 The rules on bioequiva-
lence were primarily designed to address the need
for prescribability.8 It should make little differ-
ence to a patient if the initial titration of therapy
is with a specific generic product. However, a
patient stabilised on one AED may be at risk of
that control being lost if the prescription is chan-
ged to a formulation from a different manufac-
turer. When subsequent prescriptions for the
maintenance therapy are filled, these stabilised
patients often receive a generic product from a
different supplier.
Once a number of generic products are on the
market, pharmacists may change their supplier
according to price and availability. It has been
noted by pharmacists that the lack of consistency
in supply is the main problem they experience with
generics.35 Furthermore, patients cannot usually
identify the source of a generic product (i.e. man-
ufacturer or supplier) once the product has been
dispensed so they may be unaware of such a
change,3 unless the products differ significantly
in appearance.
Economic value
Formulary committees, health policy makers,
consumer groups and others may see the increased
use of generic products as an important tool in the
battle to control healthcare costs.8 In Italy, savings
in excess of 25 million were made in 2002 as a
result of the introduction of generic drugs.1 Clearly,
there is a big incentive to cut prescribing costs,
and use of generic drugs can help. However, it is
important to ensure that the acquisition cost of the
generic product compared with the brand is not
the only cost considered. The true cost of generic
prescribing must also include the cost of additional
visits to a physician or the hospital if the substitu-
tion causes problems, and the cost of treatment
failure, if a seizure occurs. In the literature, many
commentators note that the cost of one break-
through seizure in a previously stable patient
is so high that it could offset the savings from
generics.9,12,20,27,23,32,34,36—39
170
Legal situation and informed consent
Some commentators express concern about the
medico-legal situation, if adverse consequences
arise from generic substitution.2,8,9,23,24,40 This
issue is also complicated by the concept of informed
consent. The question arises of legal responsibility
if a breakthrough seizure occurs when the medica-
tion a patient has previously received is changed for
another, considered by the regulatory authorities to
be equivalent, without the informed consent of the
patient (or the physician). While informed consent is
a prerequisite to inclusion in clinical studies, it is not
a legal obligation for switching preparations. It was
noted that patients do not reap any direct benefit
from generic substitution unless they are paying for
their medication directly, though they bear the bur-
den if any problems result. In a study of patients and
neurologists, the majority of both groups were ill-
informed about generic AEDs,24 implying that
informed consent is not currently achieved. In
another survey of physicians attending meetings on
neurology and on epilepsy, over 80% of respondents
were uncomfortable with patients receiving multiple
formulations of generic carbamazepine.40
Clinical experience
The literature review identified anecdotal reports of
problems arising from generic substitution of AEDs.
Many consisted of case reports of individuals experi-
encing adverse events (breakthrough seizures or
toxicity) on switching from branded AED to gen-
eric.14,33,39,41—43 Inevitably, such reports do not
mention the majority of patients who switched
without problem, though such studies may support
concerns about individual variation in response
being inadequately addressed under current regula-
tions. Furthermore, it is often difficult to prove that
a particular adverse event is definitely associated
with generic substitution. Case reports of problems
arising before the introduction of regulations on
bioequivalence are unlikely to be of relevance to
generic prescribing today. The literature review also
identified some crossover studies to investigate
pharmacokinetic differences between generic and
branded products.28,44—50
One study of patients treated with carbamaze-
pine, valproate or phenytoin found that 18.7% of
respondents experienced a switch in medication
supplier in the previous 2 years, and of these,
10.8% perceived problems after the switch that
were validated by their GP.12 In another retrospec-
tive survey of 81 patients, 14% reported problems
when switching from a brand to a generic product.24
P. Crawford et al.
A postal survey of 6420 neurologists found that
the majority of the 301 respondents reported pro-
blems with breakthrough seizures (68%) and/or
increased side-effects (56%) in at least one patient
with epilepsy switched from branded to generic
AED.27 Problems were also reported with switching
between generic AEDs (33% reported breakthrough
seizures in at least one patient and 27% reported
tolerability problems). Nearly half of the neurolo-
gists reporting these problems (47%) stated that two
to four patients were affected in the past year, with
over a quarter (28%) reporting breakthrough sei-
zures and/or tolerability problems with generic
AEDs in five or more patients, suggesting that con-
siderable patient morbidity occurred. As a conse-
quence, neurologists reported the need for
additional consultations (72% reported additional
phone consultations, 63% reported additional office
visits, 49% reported emergency room visits and 18%
reported hospital admissions). The impact on the
patient was also noted, with 29% reporting that the
patient missed work, and 9% reporting patient injury
as a consequence of the switched AED. Taking into
account the additional costs arising, the author
concluded that the results challenged the premise
that generic substitution of AEDs is cost-effective.
The majority of studies reporting clinical experi-
ence identified in the literature search focussed on
carbamazepine, phenytoin and valproate.
Experience with generic substitution of
carbamazepine
Carbamazepine has very low water solubility and a
narrow therapeutic range, two qualities that
increase the likelihood that generic drugs will show
therapeutic differences in the clinical situation,
even if they are considered bioequivalent.40
A number of studies have shown that there are
considerable variations in the pharmacokinetic
characteristics of different generic carbamazepine
formulations, though other studies have found few
such differences (Table 5). One study, initiated by
the FDA, found that three generic formulations were
absorbed significantly more rapidly than the inno-
vator brand.44 However, the authors concluded that
these differences would be unlikely to have clinical
effects, and that the generic formulations were
comparable to one another.
Of interest is the finding that, although statistical
similarity between products from different manu-
facturers could be demonstrated in a group as a
whole, in one study of 40 patients, a quarter had
variation in AUC greater than 20%, with differences
as high as 53%.33 A review of the literature con-
cluded that the relative bioavailability (i.e. AUC
Are there potential problems with generic substitution of antiepileptic drugs 171

Table 5 Pharmacokinetic studies on the innovator brand (Tegretol) and generic carbamazepine
References Type of study Findings
51
Single dose (8 volunteers)/multiple Increased rate of absorption of generic product
dose (5 patients) though overall extent of absorption similar
46
Cross over study in 9 healthy volunteers Five carbamazepine tablets compared.
Seven-fold differences in total bioavailability.
Central side effects (dizziness, ataxia) significantly
more common with products showing rapid absorption
38
Study in 10 patients with partial epilepsy No differences in steady state plasma levels, seizure
frequency or signs of toxicity with Tegretol and a
generic formulation
48
Double blind crossover study in 23 children Data analysed from 19 children treated with
Tegretol and generic carbamazepine for 6 weeks.
No significant differences in seizure control.
Significantly more neurological side effects with
the generic, though no apparent differences in
serum levels
47
Crossover study in 21 patients Compared multiple dose biovailability of slow
release products. Significant differences in
bioavailability (generic 11% higher than Tegretol
Retard). More seizures with branded, though the
difference was not significant
49
Double blind crossover study in 40 patients Comparison of Tegretol and a generic product.
No differences were found in pharmacokinetic
parameters or clinical efficacy
26
Cross over study in 10 patients Three formulations compared in patients already
on carbamazepine monotherapy. Little difference
in mean values of pharmacokinetic parameters,
but differences between individuals. With one
preparation, a patient showed reduction in seizure
frequency but increased toxicity
28
Cross over study in 24 healthy volunteers Wide range of bioavailability found in three
generic formulations that had been withdrawn
from use because of reports of clinical failure.
Compared with Tegretol, mean Cmax 61—74% in
two generics and 142% in one. Mean AUC varied
from 60—113% of Tegretol
45
Cross over study in 18 patients Pharmacokinetic parameters for three generic
formulations compared with Tegretol. One
generic not bioequivalent (90% CI for AUC not
within 80—120% of Tegretol)
52,53
Study in healthy volunteers Investigated criteria for assessment of
bioequivalence of controlled release
carbamazepine. Found rate of absorption
important to assess
44
Cross over study in 18 healthy volunteers Three generic formulations compared with
Tegretol. 90% CIs for Cmax for generic varied
from 111—126% and for AUC varied
from 97—108% of Tegretol. Generic products
absorbed more rapidly than Tegretol
50
Cross over study in 21 volunteers No differences in rate or extent of absorption
of two sustained release formulations of
carabamazepine
54
Study on 18 healthy volunteers Examined in vitro pharmacokinetic characteristics
of three generics and Tegretol and compared with
side effect profile. Main side effect (dizziness)
related to variation in rate of absorption
172 P. Crawford et al.

Table 6 Case reports of problems following replacement of the innovator brand (Tegretol) with generic carbama-
zepine
References Type of study Findings
41
Case report of 16-year old boy Boy with partial epilepsy caused by cerebral hemiatrophy
stable on Tegretol experienced convulsions when switched
to the generic
14
Case report of three patients Loss of seizure control following generic substitution of Tegretol,
with restoration of control when Tegretol reinstituted
42
Case report of woman Seizure activity increased following generic substitution of
Tegretol, with fall in serum carbamazepine level. Control
regained when Tegretol reinstituted
43
Case report of two patients Breakthrough seizures occurred within 3—7 days of generic
substitution of Tegretol
39
Case report of two patients Breakthrough seizures associated with drop in serum levels of
carbamazepine following switch from Tegretol to generic
33
Case reports of two Increases in carbamazepine Cmax of 22% and 41% after mandatory
6-year-old children generic substitution, resulting in toxicity that reversed when
Tegretol reinstituted. One child required hospitalisation

generic drug/AUC brand name drug) varied widely
both between generic formulations and between
batches within the same formulation.34
Individual variations in pharmacokinetic para-
meters may explain the case reports of break-
through seizures on generic substitution (Table 6).
Indeed, it has been estimated that over 20% of cases
of loss of efficacy with carbamazepine may be
traced to lowering of serum levels following a switch
to a generic product.55 There have also been reports
of increased tolerability problems with generic sub-
stitution.34 Following the replacement of the inno-
vator brand, Tegretol, with generic products, there
have been reports of a higher incidence of neuro-
logical side effects and skin rash.34
The potential risks to patients from substituting
Tegretol result in few cost savings,34 estimated at
less than the additional costs incurred as a result.39
In the case of a child hospitalised because of toxicity
arising from substitution of generic carbamazepine
for the brand, the cost of $8000 was noted to greatly
overshadow the $3.25 weekly savings from generic
substitution,33 though the frequency of such
adverse consequences was not assessed.
Experience with generic substitution of
phenytoin
Phenytoin has low water solubility, narrow thera-
peutic range and nonlinear kinetics–—all risk factors
for potential problems with generic substitution.23
Many papers in the 1960s and 1970s reported
differences in pharmacokinetics between the inno-
vator brand (Dilantin) and generic phenytoin,56—62
and clinical problems from breakthrough seizures and
toxicity.63—65 Many of these reports originated in
Australia and were caused by a change in excipient.31
More recent studies on bioequivalence have
reported differences in pharmacokinetic para-
meters between generic phenytoin and the innova-
tor brand,30,66,67 though some of these differences
were small and considered unlikely to have clinical
consequences.29,68 Differences in the content of
phenytoin between capsules of generic and branded
preparations, rather than differences in the phar-
macokinetics per se, have also been reported.31
Differences in pharmacokinetic parameters
between branded and generic formulations have
been reported in Chinese subjects69 and when taken
with food,16 which could have clinical implications.
Experience with generic substitution of
valproate
Valproate has a relatively narrow therapeutic range
and nonlinear pharmacokinetics–—both risk factors
for difficulties with generic substitution.23
In an open study, 64 patients with seizure dis-
orders and mental retardation were randomly
assigned to generic valproate, or continued to
receive the branded innovator product (Depakene)
for 4 weeks, and were then switched to the other
therapy.70 There were no significant changes in
blood levels of the products, or in seizure frequency,
though the same manufacturer produced the cap-
sules for both the branded and the generic compa-
nies. The authors concluded that switching from the
branded product (priced at $94.23 for 100 capsules
of 250 mg) to the generic product (priced at $7.39)
would result in considerable cost savings.
However, there have also been at least two
reports of patients experiencing problems when
switching between the branded and generic pro-
duct. In one case, a 19 year old woman experienced
Are there potential problems with generic substitution of antiepileptic drugs 173

Table 7 Recommendations on generic substitution from the American Academy of Neurology9

Generic substitution can be approved only if safety and efficacy are not compromised
Physicians should avoid switching between formulations of AEDs
Specific information about each AED generic (e.g. AUC, Cmax, reported complications)
should be made available to physicians
Pharmacists should be required to inform patients and physicians when switching a patient between
manufacturers, with labelling to allow identification of specific manufacturers
Organisations that encourage or mandate substitution of AEDs (e.g. federal or state agencies, health plans, hospitals)
should evaluate their responsibility for problems arising from the policy
Further research on the impact of generic substitution is required

Table 8 Special categories proposed for exemption from mandatory generic substitution2
Special categories Examples
Critical patients Very old, very young, those suffering from multiple diseases
who are managed with multiple drugs, those living alone
Critical diseases Intercurrent disorders, mainly in chronic care, where
drug-disease interactions present major problems, those with
serious clinical sequelae, should therapy fail
Critical drugs Narrow therapeutic index, those requiring a complex therapeutic
regimen, those with considerable drug interactions, those requiring
titration of individual doses

breakthrough seizure on the switch to the generic,17
while in the other report, a 44 year old woman
suffered gastrointestinal side effects after switch-
ing from an enteric-coated branded tablet to a
generic formulation.18
Implications
Although the evidence is limited, the substantial
concerns expressed about potential problems aris-
ing from uncontrolled switching of patients between
older AEDs from different manufacturers appear to
have some grounding in clinical experience. These
findings suggest that, although mean values for
bioavailability for licensed generic products are
within the range specified by the authorities, varia-
bility in response means that some individuals (par-
ticularly those with well-controlled epilepsy) may
be placed at unjustified risk by generic substitu-
tion.9 The current recommendations of the Amer-
ican Academy of Neurology9 are summarised in
Table 7, though these are under review.
It has been suggested that generic substitution is
not allowed where there are special factors that
could result in therapeutic differences, even if
products are considered by regulatory authorities
as bioequivalent (Table 8).
In recognition of these issues, the FDA has listed
carbamazepine, phenytoin and valproate as pro-
ducts with restrictions on generic substitution.8 In
Spain, carbamazepine and gabapentin are excluded
from a reference list, despite generics being avail-
able, indicating that pharmacists are not expected
to substitute the generic for the branded product. In
Germany, a law for reduction of drug related costs in
healthcare became effective in February 2002,
which allows pharmacists to substitute a generic
product instead of the brand, unless explicitly pro-
hibited by the physician. However, the German
Section of the International League Against Epilepsy
published a statement against this approach for
AEDs (Table 9),7 and AEDs were not included in
the final list of drugs considered as suitable for
substitution. In Denmark, some AEDs containing
oxcarbamazepine are exempt from generic substi-
tution because of bioequivalence problems, while
Finland has exempt all AEDs from substitution.
Table 9 Objections to uncontrolled generic substitu-
tion of AEDs7
The consequences of therapeutic differences are
great–—one breakthrough seizure can result in
loss of a job or a driving licence
Permitted differences in bioavailability of generic
products compared with the brand may have
therapeutic implications
Sustained release and immediate release formulations
of an AED are not interchangeable
Patients may have concerns that generic substitution
could lead to worsening control
Seizure-free patients should not have their therapy
altered
174
South Africa recommends that cabamazepine and
phenytoin are not substituted, while noting that it is
also not advisable for any medicines with narrow
therapeutic range, erratic intra/inter patient
responses, dosage forms likely to result in bioavail-
ability problems (e.g. delayed release) and for use in
critically ill, paediatric or geriatric populations. In
the UK, brand name prescribing for AEDs is recom-
mended.
As other AEDs lose patent protection, it is impor-
tant to ensure that patients, pharmacists, prescri-
bers and decision-makers are all aware of the issues
to consider. Although newer AEDs, such as lamotri-
gine, gabapentin, topiramate and levetriacetam,
are not considered to have a narrow therapeutic
range like carbamazepine or valproate, some of the
issues of concern still apply.24 Like the older AEDs,
they require individual titration, and the conse-
quences of a breakthrough seizure are the same,
regardless of the therapy that failed. In the light of
the concerns expressed and clinical experience with
older AEDs, more evidence is required before gen-
eric substitution of other AEDs is permitted. This has
recently been recognised in Sweden, which has
added gabapentin to the list of products like carba-
mazepine and valproate already exempt from man-
datory generic substitution. Furthermore, the legal
situation should be clarified, with responsibilities
assigned for ensuring patients and physicians are
informed.
Conclusions
Generic prescribing is an important tool in ensuring
effective use of limited healthcare resources. While
this approach is widely supported, the concerns
about uncontrolled switching of AED manufacturers
expressed by physicians caring for patients with
epilepsy must be noted:
 potentially serious consequences of any failure of
therapy;
 potential for adverse events and variability of
response to certain AEDs in some individuals;
 need for careful titration of AEDs and complex
management regimens that have a high potential
for disruption if any component is changed in
certain patients;
 bioequivalence, as defined in regulations, does
not always correspond to therapeutic equivalence
for AEDs, because of the permitted range, eva-
luation methods and individual variation;
 high potential for repeated switching of AEDs
from different manufacturers due to poor conti-
nuity of supply and pharmacists’ cost concerns;
P. Crawford et al.
 savings made from generic prescribing of AEDs
may be outweighed by the cost of adverse con-
sequences in some patients;
 potential medico-legal consequences if adverse
consequences arise in a patient who did not
give informed consent to switching of AED
manufacturer.
Clinical experience appears to support these con-
cerns, which has resulted in restrictions on use of
specific generic AEDs in some countries. Caution is
advised on generic substitution for newly off-patent
AEDs. The challenge is to gather more data to
identify groups and drugs (especially among the
newer AEDs) which pose the highest risk and to
determine whether there are other factors, apart
from bioavailability variations, that account for the
lack of therapeutic equivalence among different
preparations in some individuals. Studies to address
these issues comprehensively with all of the newer
agents are unlikely to be conducted. Therefore it is
prudent for patients, neurologists and pharmacists
to be aware of the issues and to approve generic
prescribing of AEDs for certain high-risk patients
prior to it being instituted.
References
1. Borgerhini G. The bioequivalence and therapeutic efficacy of
generic versus brand-name psychoactive drugs. Clin Ther
2003;25:1578—92.
2. Lamy PP. Generic equivalents: issues and concerns. J Clin
Pharmacol 1986;26:309—16.
3. Anonymous. For and against generic prescribing. Drug Ther
Bull 1987;25:93—5.
4. Richens A. Impact of generic substitution of anticonvulsants
on the treatment of epilepsy. CNS Drugs 1997;8:124—33.
5. Besag FMC. Is generic prescribing acceptable in epilepsy?
Drug Safety 2000;23:173—82.
6. Meredith P. Bioequivalence and other unresolved issues in
generic substitution. Clin Ther 2003;25:2875—90.
7. Krämer G, Schneble H, Wolf P. Risks of the new ‘‘aut idem’’
regulations for treatment with antiepileptic drugs. Akt Neu-
rol 2002;29:115—22.
8. Banahan BF, Bonnarens JK, Bentley JP. Generic substitution of
NTI drugs: issues for formulary committee consideration.
Formulary 1998;33:1082—96.
9. American Academy of Neurology. Assessment: generic sub-
stitution for antiepileptic medication. Neurology 1990;40:
1641—3.
10. EUCARE. European white paper on epilepsy. Epilepsia
2003;44(Suppl. 6):1—88.
11. Jacoby A, Baker G, Steen N, Potts P, Chadwick D. The clinical
course of epilepsy and its psychosocial correlates: findings
from a UK community study. Epilepsia 1996;37:148—61.
12. Crawford P, Hall WW, Chappell B, et al. Generic prescribing
for epilepsy. Is it safe? Seizure 1996;5:1—5.
13. Blier P. Brand versus generic medications: the money, the
patient and the research. J Psychiatry Neurosci 2003;28:
167—8.
Are there potential problems with generic substitution of antiepileptic drugs
14. Sachdeo RC, Belendiuk G. Generic versus branded carbama-
zepine. Lancet 1987;I:1432.
15. Meyer MC, Straughn AB. Biopharmaceutical factors in seizure
control and drug toxicity. Am J Hosp Pharm 1993;50(Suppl.
5):S17—22.
16. Wilder BJ, Leppik I, Hietpas TJ, et al. Effect of food on
absorption of Dilantin Kapseals and Mylan extended pheny-
toin sodium capsules. Neurology 2001;57:582—9.
17. MacDonald JT. Breakthrough seizure following substitution of
Depakene capsules (Abbott) with a generic product. Neurol-
ogy 1987;37:1885.
18. Brown ES, Shellhorn E, Suppes T. Gastrointestinal side effects
after switch to generic valproic acid. Pharmacopsychiatria
1998;31:114.
19. Wyllie E, Pippenger CE, Rothner AD. Increased seizure fre-
quency with generic primidone. J Am Med Assoc 1987;258:
1216—7.
20. Jobst BC, Holmes GL. Prescribing antiepileptic drugs. Should
patients be switched on the basis of costs? CNS Drugs
2004;18:617—28.
21. Benet LZ, Goyan JE. Bioequivalence and narrow therapeutic
index drugs. Pharmacotherapy 1995;15:433—40.
22. FDA Code of Federal Regulations. Washington, DC: US Gov-
ernment Printing Office; 1997. 21CFR320.33(c).
23. Nuwer MR, Browne TR, Dodson WE, et al. Generic sub-
stitutions for antiepileptic drugs. Neurology 1990;40:
1647—51.
24. Guberman A, Corman C. Generic substitution for brand name
antiepileptic drugs: a survey. Can J Neurol Sci 2000;27:37—
43.
25. European Commission. Marketing authorisation. November
2002.
26. Wolf P, May T, Tiska G, Schreiber G. Stead sate concentrations
and diurnal fluctuations of carbamazepine in patients with
different slow release formulations. Arzneimittel For-
schung/Drug Res 1992;42:284—8.
27. Wilner AN. Therapeutic equivalency of generic antiepileptic
drugs: results of a survey. Epilepsy Behav 2004;5:995—8.
28. Meyer MC, Straughn AB, Jarvi EJ, et al. The bioinequivalence
of carbamazepine tablets with a history of clinical failures.
Pharm Res 1992;9:1612—6.
29. Hirji MR, Measuria H, Kuhn S, Mucklow JC. A comparative
study of the bioavailability of five different phenytoin pre-
parations. J Pharm Pharmacol 1985;37:570—2.
30. Rosenbaum DH, Rowan AJ, Tuchman L, French JA. Compara-
tive bioavailability of a generic phenytoin and Dilantin.
Epilepsia 1994;35:656—60.
31. Soryal I, Richens A. Bioavailability and dissolution of proprie-
tary and generic formulations of phenytoin. J Neurol Neu-
rosurg Psychol 1992;55:688—91.
32. Meredith PA. Generic drugs: therapeutic equivalence. Drug
Safety 1996;15:233—42.
33. Gilman JT, Alvarez LA, Duchowny M. Carbamazepine toxicity
resulting form generic substitution. Neurology 1993;43:
2696—7.
34. Brodie MJ, Johnson FN. Carbamazepine in the treatment
of seizure disorders: efficacy, pharmacokinetics and adverse
event profile. Rev Contem Pharmacother 1997;8:87—122.
35. Maline N. What pharmacists really think about brand sub-
stitutions. Drug Store News Pharm 1996;6:23—30.
36. Heaney DC, Beran RG, Halpern MT. Economics in epilepsy
treatment choices: our certain fate? Epilepsia 2002;43(Suppl.
4):32—8.
37. Heaney DC, Begley CE. Economic evaluation of epilepsy treat-
ment: a review of the literature. Epilepsia 2002;43(Suppl.
4):10—7.
175
38. Jumo-as A, Bella I, Craig B, Lowe J, Dasheiff RM. Comparison
of steady state blood levels of two carbamazepine formula-
tions. Epilepsia 1989;30:67—70.
39. Welty TE, Pickering PR, Hale BC, Arazi R. Loss of seizure
control associated with generic substitution of carbamaze-
pine. Ann Pharmacother 1992;26:775—7.
40. Wilner AN. Physicians underestimate the frequency of gen-
eric carbamazepine substitution: results of a survey and
review of the problem. Epilepsy Behav 2002;3:522—5.
41. Pedersen SA, Dam M. Karbamazepin: er synonympräparater
ens. Ugeskr Laeg 1985;147:2676—7.
42. Koch G, Allen JP. Untoward effects of generic carbamazepine
therapy. Arch Neurol 1978;44:578—9.
43. Hartley R, Aleksandrowicz J, Ng PC, et al. Dissolution and
relative bioavailability of two carbamazepine preparations for
children with epilepsy. J Pharm Pharmacol 1991;43:117—9.
44. Meyer MC, Straughn AB, Nhatre RM, et al. The relative
bioavailability and in vivo—in vitro correlations for four
marketed carbamazepine tablets. Pharm Res 1998;15:
1787—91.
45. Silpakit O, Amornpichetkoon M, Kaojarern S. Comparative
study of bioavailability and clinical efficacy of carbamaze-
pine in epileptic patients. Ann Pharmacother 1997;31:548—
52.
46. Neuvonen PJ. Bioavailability and central side effects of
different carbamazepine tablets. Int J Clin Pharmacol Ther
Toxicol 1985;23:226—332.
47. Reunanen M, Heinonen EH, Nyman L, Anttila M. Comparative
bioavailability of carbamazepine from two slow-release pre-
parations. Epilepsy Res 1992;11:61—6.
48. Hartley R, Aleksandrowicz J, Ng PC, et al. Breakthrough
seizures with generic carbamazepine: a consequence of
poorer bioavailability? Br J Clin Practice 1990;44:270—3.
49. Oles KS, Penry JK, Smith LD, et al. Therapeutic bioequiva-
lency study of brand name versus generic carbamazepine.
Neurology 1992;42:1147—53.
50. Wangemann M, Retzow A, Evers G, et al. Bioavailability study
of two carbamazepine containing sustained release formula-
tions after multiple oral dose administration. Arzneimittel
Forschung/Drug Res 1998;48:1131—7.
51. Glende M, Hüller H, Mai I, et al. Comparative bioavailability
of two carbamazepine tablets. Int J Clin Pharmacol Ther
Toxicol 1983;21:631—3.
52. Bialer M, Yacobi A, Moros D, et al. Criteria to assess in vivo
performance and bioequivalence of generic controlled-
release formulations of carbamazepine. Epilepsia 1998;39:
513—9.
53. Bialer M, Arcavi L, Sussan S, et al. Existing and new criteria
for bioequivalence evaluation of new controlled release (CR)
products of carbamazepine. Epilepsy Research 1998;32:
371—8.
54. Olling M, Mensinga TT, Barends DM, et al. Bioavailability of
carbamazepine from four different products and the occur-
rence of side effects. Biopharm Drug Dispos 1999;20:19—28.
55. Jain KK. Investigation and management of loss of efficacy of
an antiepileptic medication using carbamazepine as an
example. J R Soc Med 1993;86:133—6.
56. Appleton DB, Eadie MJ, Hooper WD, et al. Blood phenytoin
concentrations produced by ingestion of there different
phenytoin preparations. Med J Aust 1972;1:410—2.
57. Lund L. Clinical significance of generic inequivalence of three
different pharmaceutical preparations of phenytoin. Eur J
Clin Pharmacol 1974;7:119—24.
58. Sansom LN, O’Reilly WJ, Wiseman CW, Stern LM, Derham J.
Plasma phenytoin levels produced by various phenytoin pre-
parations. Med J Aust 1975;2:593—5.
176
59. Pentikäinen PJ, Neuvonen PJ, Elfving SM. Bioavailability of
four brands of phenytoin tablets. Eur J Clin Pharm 1975;
9:213—8.
60. Stewart MJ, Ballinger BR, Devlin EJ, Miller AY, Ramsay AC.
Bioavailability of phenytoin. A comparison of two prepara-
tions. Eur J Clin Pharm 1975;9:209—12.
61. Tammisto P, Kauko K, Viukari M. Bioavailability of phenytoin.
Lancet 1976:254—5.
62. Rambeck B, Boenigk H-E, Stenzel E. Bioavailability of three
phenytoin preparations in healthy subjects and in epileptics.
Eur J Clin Pharm 1977;12:285—90.
63. Rail L. Dilantin overdosage. Med J Aust 1968:339.
64. Eadie MJ, Sutherland JM, Tyrer JH. Dilantin overdosage. Med
J Aust 1968:515.
65. Tyrer JH, Eadie MJ, Sutherland JM. Investigation of an out-
break of anticonvulsant intoxication. Proc Aust Assoc Neurol
1970;7:15—8.
P. Crawford et al.
66. Hodges S, Forsythe WI, Gillies D, Remington H, Cawood A.
Bioavailability and dissolution of three phenytoin pre-
parations for children. Dev Med Child Neurol 1986;28:
708—12.
67. Mikati M, Bassett N, Schachter S. Double-blind randomised
study comparing brand name and generic phenytoin mono-
therapy. Epilepsia 1992;33:359—65.
68. Chen S-S, Allen J, Oxley J, Richens A. Comparative bioavail-
ability of phenytoin from generic formulations in the United
Kingdom. Epilepsia 1982;23:149—52.
69. Tsai J-J, Lai M-L, Yang Y-HK, Huang J-D. Comparison of
bioequivalence of four phenytoin preparations in patients
with multiple-dose treatment. J Clin Pharmacol 1992;32:
272—6.
70. Vadney VJ, Kraushaar KW. Effects of switching from Depa-
kene to generic valproic acid on individuals with mental
retardation. Mental Retard 1997;35:468—72.