Current Trends in Antiepileptic Drug Therapy


Address correspondence and reprint requests to Dr. E. Perucca at Clinical Pharmacology Unit, Department of Internal Medicine and Therapeutics, University of Pavia, Piazza Botta 10, 27100 Pavia, Italy.


Summary:  Over the last two decades, drug therapy for epilepsy has improved substantially. This can be ascribed to a large extent to three factors, including the demonstration of the advantages of monotherapy; the realization of the need for dosage tailoring, coupled [for some antiepileptic drugs (AEDs)] with control of pharmacokinetic variability through therapeutic drug monitoring; and the introduction of newer agents with improved tolerability profiles. What further advances should we expect for the future? Current trends that are expected to increasingly affect our prescribing patterns include greater reliance on evidence-based medicine and treatment guidelines, a trend that will be facilitated by completion of therapeutically meaningful randomized trials (including cost-effectiveness studies) and high-quality observational studies (including multinational pregnancy registries), as well as initiatives from scientific societies and government organizations aimed at condensing the most relevant information into therapeutic guidelines. The explosion in communication technology will accelerate dissemination of this information and its application to clinical practice. Other factors include a more rational patient-tailored AED selection and dose individualization, aided by characterization of predictors of outcome as defined by clinical parameters (sex, age, epilepsy syndrome, and etiology), pathophysiological mechanisms, and newly discovered genetic markers of outcome; improved definition of the role of new AEDs, resulting in their increased use in newly diagnosed epilepsy; and reappraisal of the value of combination therapy in refractory epilepsies, based on evidence produced by experimental and clinical studies designed to identify favorable pharmacodynamic interactions. Additional important developments may come from the discovery of novel, more efficacious AEDs and from exploration of potential new targets, such as prevention of epileptogenesis.

Over the last 25 years, the pharmacological management of the epilepsies has made tremendous progress, resulting in an increasing proportion of patients achieving seizure freedom without undue adverse effects (1). A number of milestones have contributed to these advances. First of all, recognition of the value of monotherapy has been seminal in exploiting the therapeutic properties of individual antiepileptic drugs (AEDs) and in avoiding the acute and chronic toxicity associated with unnecessary polytherapy (2,3). At the same time, realization that AEDs exhibit a large pharmacokinetic and pharmacodynamic variability has led to widespread recognition that adequate responses can be achieved only with careful individualization of dosage, aided in some cases by measurements of serum drug concentrations (4). Another major step forward came with the increasing application of randomized controlled trials designed to evaluate the comparative efficacy and tolerability of AEDs. In particular, studies such as the first Veterans Administration (VA) Cooperative Study, which showed the superior effectiveness of phenytoin and carbamazepine (CBZ) over barbiturates in patients with localization-related epilepsy (5), have had a tremendous impact on general prescribing. Last, the advent of a novel generation of AEDs has greatly expanded the therapeutic armamentarium, improving our ability to tailor drug choice to the characteristics of the individual patient (6).


Despite the major advances previously summarized, the current approach to AED therapy still has many limitations. For example, inadequate understanding of the pathophysiological mechanisms underlying seizure disorders, coupled with the still incomplete knowledge of the mechanisms of action of individual AEDs, prevents a mechanistic approach to AED therapy. Indeed, drug selection in epilepsy remains based on empirical assessment of the probability of achieving seizure freedom and associated side-effect profiles, rather than on the rational application of a treatment that corrects a specific functional or biochemical abnormality (7,8). For the same reasons, we have no reliable tools to predict which AED, among the many known to be active against a given seizure type, will control the seizures in an individual patient. Therefore, at the current state of knowledge, AED therapy is based mainly on a trial-and-error approach.

Clearly, however, the most important limitation of current pharmacological treatments is that, in at least one third of patients, seizures cannot be controlled completely by available AEDs (9,10). The phenomenon of drug resistance has hardly been affected by the advent of second-generation AEDs, as no more than 10–15% of patients with severe refractory epilepsy become seizure free by treatment with these agents (6,11). Therefore, the search for newer and more effective AEDs should continue.

The purpose of this article is to provide a concise overview of a number of strategies that are currently being implemented in an attempt to overcome the limitations discussed previously and further improve the effectiveness of AED therapy. Five such strategies will be discussed in some detail, including (a) an increasing reliance on therapeutic guidelines and evidence-based medicine; (b) a more rational patient-tailored selection of AEDs and drug dosages; (c) a better definition of the indications of new AEDs; (d) novel approaches to AED combinations; and (e) the development of newer AEDs, novel modes of administration, and novel therapeutic targets (Table 1).

Table 1. Trends in the pharmacological treatment of epilepsy
  1. AED, antiepileptic drug.

Increased reliance on therapeutic guidelines and evidence-based medicine
Improved tailoring of AED choice and dosages
Better definition of new AED indications
Novel approaches to AED combinations
Development of novel AEDs, novel modes of administration, and novel therapeutic targets


As discussed previously, a fully rational approach to the treatment of epilepsy directed at the underlying etiology is not currently possible, except in special circumstances (e.g., the use of pyridoxine to suppress neonatal seizures secondary to pyridoxine deficiency). Therefore, therapeutic decisions have been traditionally based on the physician's subjective perception of the pros and cons of individual AEDs, personal experience, and marketing pressure, rather than on sound evaluation of scientific evidence concerning the relative merits of available treatments (7,8). However, this situation is rapidly changing due to a number of factors.

First of all, in the last 15 years there has been an explosion of prospective studies using the most powerful scientific tool in medicine, the randomized controlled trial (RCT). Admittedly, most of these trials were designed to address regulatory needs for the licensing of a new product; as such they rarely provide the kind of information that is most important for rational prescribing (12,13). For example, most RCTs of new AEDs as adjunctive therapy were designed simply to show that the test drug was superior to placebo. This is indeed essential to determine whether a drug should be granted marketing authorization. However, it does not really address the concerns of the treating physician, who foremost wants to know how that agent compares with other AEDs in terms of efficacy and tolerability. Even active-control monotherapy trials may fail to provide this information because the experimental design of these trials often involves suboptimal duration of follow-up, as well as dosages, titration rates, or dosing schedules that are biased in favor of the sponsor's product (14). While these limitations should be understood, it should be acknowledged that RCTs provide, to a variable extent, valuable information that, when interpreted correctly, can be usefully applied to clinical practice. Moreover, many RCTs performed in recent years have addressed important issues other than the simple testing of a new agent for regulatory purposes. Examples include (a) the demonstration that AEDs given after a first unprovoked tonic-clonic seizure reduce the risk of a recurrence (15), even though the risk-to-benefit ratio is usually against prescribing drug therapy in these patients (particularly because early initiation of therapy does not appear to improve the long-term prognosis) (16); (b) the finding that systematic monitoring of the serum concentrations of certain AEDs does not necessarily improve clinical outcome compared with dosage adjustment on clinical grounds (17); (c) the demonstration that long-term pharmacological prophylaxis is ineffective in preventing the development of chronic epilepsy following febrile seizures, supratentorial surgery, or brain trauma (18); and (d) the quantitation of the relative risks of seizure recurrence following discontinuation of AED therapy in patients who had been free from seizures for 2 years or longer (19). It is clear that the results of many of these trials are extremely relevant for informed decisions in everyday clinical practice, and in some cases, as in the use of long-term AED prophylaxis after neurosurgery or head trauma, they expose the inappropriateness of prescribing patterns that had been consolidated for decades (20,21). Additionally, RCTs provide a constant reminder of the pitfalls involved in making noncritical inferences in clinical medicine. For example, when a patient has seizure relapse shortly after discontinuing a previously effective treatment, it is customary to ascribe this event to the lack of therapeutic coverage afforded by the drugs. However, a well-designed RCT (22) elegantly demonstrated that in many cases other factors may be at play, as indicated by the fact that seizure relapse, while more frequent after drug withdrawal, is also relatively common in patients randomized to remain on treatment (Fig. 1). The importance of these observations in providing an evidence-based assessment of risk-to-benefit ratios when contemplating discontinuation of therapy is obvious.

Figure 1.

Randomized, controlled trial on the effect of withdrawing AED therapy in patients with epilepsy who had been seizure free for at least 2 years. Patients were randomized to continue their treatment (controls) or to have their drug treatment gradually withdrawn. The illustration shows the actuarial percentage of patients remaining free from seizures after randomization in the two groups. Reproduced with permission from The Medical Research Council Antiepileptic Drug Withdrawal Study Group. Lancet 1991; 337:1175–80.

Careful weighing of available scientific evidence as a basis for informed decisions is an accelerating trend in everyday practice. In addition to recognition of the importance of a sound methodological approach in clinical practice, two factors are contributing to this trend. The first is the development of therapeutic guidelines and protocols by health authorities, scientific societies, and nongovernmental organizations, all of which provide the prescriber with accessible and critically assessed practical information. Organizations actively involved in the production of guidelines on various aspects of epilepsy therapy include the ad hoc subcommittees of the American Academy of Neurology, the American Academy of Pediatrics, and the American Epilepsy Society (23–25), the Scottish Intercollegiate Guidelines Network (26), the U.K. National Institute of Clinical Excellence (27), and the International League Against Epilepsy (International League Against Epilepsy, Commission on Therapeutic Strategies, Subcommission on AED Guidelines. Guidelines on the pharmacological treatment of children and adults with epilepsy, unpublished report). The procedures for the development and validation of therapeutic guidelines are well codified (28), the aim being to provide the physician with information about the level of scientific evidence on which any recommendation is based. Advantages (and aims) of well-designed guidelines should include the dissemination of high educational standards among medical practitioners, a rational approach to the clinical decision process, a minimization of commercial influences on prescribing, a better allocation of health resources due to the application of cost-effective interventions, a reduced liability to litigation resulting from inappropriate treatments, and, ultimately, a better health outcome for patients. It is clear, however, that guidelines cannot be substituted for competent individualized clinical judgment, as each patient may exhibit unique features not addressed by a general guideline. Likewise, guidelines are not a substitute for the physician's duty to remain independently informed of advances in medical knowledge.

The second factor that facilitates application of an evidence-based approach is the explosion in communication technology. Continuing medical education (CME) activities are expanding at an unprecedented pace. Equally important are the search engines, literature databases, and online journals that are now easily accessible to most health professionals, even in remote parts of the world. Indeed, the difficulty is no longer to be able to access information, but to be able to distinguish what is really relevant from what is not. It is at this level that meta-analytic studies, systematic reviews, high-quality CME activities, and therapeutic guidelines can make the difference. If keeping abreast of the overflow of information can prove to be a disorienting challenge for physicians, it is much more so for laypersons and patients, who increasingly come to us already equipped with a wealth of information, or misinformation, about the nature of their symptoms and potential treatments. A recent search on the Google search engine identified 738,000 links for the term “epilepsy” (T. Tomson, October 2002, personal communication). The number of links for each of the new AEDs using Google compared with PubMed (Fig. 2) raises intriguing questions about how patients and physicians can handle currently available sources of information. This is a new challenge that the medical profession should be ready to address.

Figure 2.

The Internet as a source of information on second-generation AEDs. A number of links were retrieved for each drug using the Google search engine compared with PubMed. The search was conducted in October 2002 (courtesy of Dr. T. Tomson, Stockholm).


Compared with other disorders, epilepsies exhibit extreme heterogeneity in underlying etiologies, pathophysiological mechanisms, and clinical manifestations. Almost all RCTs of existing AEDs have selected patients on the basis of seizure type, and little or no effort has been made to explore comparative drug effects in patients with homogeneous clinical characteristics (12). This is regrettable because differences in underlying pathophysiological mechanisms could explain interindividual differences in drug response, even among patients with the same syndrome. For example, etiology has been little investigated as a factor affecting differential responses to AEDs, yet it may be an important predictor, as most clearly documented by the finding that vigabatrin (VGB) is more efficacious in infantile spasms associated with tuberous sclerosis than in spasms associated with other etiologies (29,30). Elucidation of the mechanisms responsible for seizure generation in specific syndromes and subsyndromes could be important in explaining and predicting drug responses. For example, the discovery of mutations in the sodium channel SCN1A gene in patients with severe myoclonic epilepsy of infancy (31) may provide an explanation for the paradoxical aggravation of seizures caused by lamotrigine (LTG), a sodium channel blocker, in these patients (32). Ultimately, improvement in our understanding of how individual AEDs interfere with the molecular mechanisms underlying specific forms of epilepsy will allow us to tailor drug choice to maximize efficacy in the individual patient. Pharmacogenomic information could also provide important clues (33). Similar considerations apply to the prediction of adverse effects, for example, through identification of patients who carry a genetically determined susceptibility to idiosyncratic drug reactions (33). Again, this should foster patient-tailored AED therapy for individual patients by allowing avoidance of compounds with unacceptable toxicity risks.

Evidence-based, patient-tailored AED therapy is going to be more frequently applied in the future thanks to the increasing availability of RCTs in specific patient groups. For example, epilepsy in the elderly differs in many ways from that seen in younger age groups, and response to AEDs may also be altered in old age, making it difficult to extrapolate results from studies performed in nonelderly populations (34). The VA Cooperative Study (35), which is nearing completion, is comparing a variety of AEDs in patients with onset of epilepsy in old age; its results are likely to provide valuable data for rational drug selection in this growing segment of the population.

Many people with epilepsy suffer from comorbidities, which may be favorably or adversely affected by specific AEDs. Again, this is an important consideration in tailoring choice and dosage of AEDs to the needs of the individual patient. Several studies in recent years have demonstrated that some AEDs are effective in nonepilepsy indications, an observation that can be exploited to a patient's benefit. For example, gabapentin would be a natural choice to treat a patient with diabetes who suffers at the same time from epilepsy and neuropathic pain (36), whereas valproate (VPA) would allow the simultaneous management of seizures and migraine (37), and LTG would be valuable to manage the comorbidity of epilepsy and bipolar depression (38). As more information emerges from RCTs of AEDs in an array of potential additional indications, these considerations will carry increasing weight in therapeutic decisions.


Second-generation AEDs have been initially investigated in placebo-controlled adjunctive therapy trials in adults with refractory partial epilepsy (6). More recently, however, RCTs have been extended to other syndromes, special age groups (such as infants, children, and the elderly), and monotherapy indications (13). Newer drugs are increasingly being compared with older-generation agents (14), and even head-to-head comparisons of new-generation AEDs are being completed (39). Moreover, well over one million patients have been exposed to some recently developed AEDs, allowing collection of important data concerning safety and potential efficacy in special patient groups. All this information helps to better define the role of these drugs in the treatment algorithm.

Initially, new AEDs were only used as add-on therapy in highly refractory patients (40). However, as knowledge and experience expand, physicians feel increasingly confident in including newer AEDs among the options to be considered at the very onset of therapy (6). Each of the currently available drugs differs from others in terms of spectrum of activity, side-effect profile, ease of use, interaction potential, and cost. Therefore, each could offer unique advantages in specific situations. There are already examples of patient groups in which use of a new AED can be proposed as first-line therapy. One is the use of VGB in the de novo management of infantile spasms, particularly in infants with tuberous sclerosis (30). In this situation, benefits are likely to outweigh the risk of retinal toxicity in view of the limited duration of drug exposure and the potential toxicity associated with alternative treatments such as steroids and adrenocorticotropic hormone (7). Likewise, LTG has been proposed for the first-line treatment of epilepsy with onset in old age because of its particularly favorable tolerability profile in this age group (41). Increased use of second-generation AEDs in the future may also be facilitated by a relative decline in their costs, partly related, in some cases, to the introduction of generic products.

Despite extensive research, we still lack critical information for rational prescribing in many areas. The most important case is the lack of reliable data on relative risks associated with the use of old- and new-generation AEDs during pregnancy (42). Multinational prospective collaborative registries have been set up to compare the incidence of major malformations in the offspring of women exposed to different AEDs and AED combinations during pregnancy (43,44). Ongoing studies are also assessing the effects of prenatal AED exposure on postnatal development, with special emphasis on cognitive functions. Ultimately, the results of these studies could revolutionize AED prescribing for women with childbearing potential.


While everyone acknowledges the advantages of monotherapy (2,3), there is no doubt that in some patients with difficult-to-treat epilepsy only a combination of two or more AEDs allows achievement of seizure freedom or, at least, the best compromise between seizure control and adverse effects (1,7). The term “rational polypharmacy” has been used to describe the appropriate application of multiple drug therapy in patients whose seizures cannot be controlled with a single drug (45). This term has also been applied to imply that AEDs can be combined rationally based on knowledge of their mechanisms of action and/or clinical activity profiles. In fact, the process by which we select and use drug combinations is far from rational (45). Although the suggestion has been made that there are advantages in combining drugs with different or “complementary” modes of action (46), in practice this approach has not been validated in the clinical setting. Its applicability may be negated by the fact that almost all AEDs have multiple modes of action and many of these actions are still poorly understood (47,48). While these limitations are acknowledged, empirical evidence indicates that certain AED combinations offer a superior therapeutic index compared with others. For example, undesired pharmacodynamic interactions leading to potentiation of central nervous system adverse effects have been described in patients given a combination of oxcarbazepine and CBZ (49), or LTG and CBZ (50). Conversely, a combination of LTG and VPA, used at appropriately adjusted dosages, could be particularly beneficial in patients whose seizures were not controlled by either drug given alone (51,52), even though concern has been generated by preliminary findings suggesting a higher teratogenicity risk after exposure to this combination (53,54). Identification of the relative merits (and risks) of different AED combinations is a priority in current research, and it is likely to increasingly influence therapeutic decisions in the future.


The fact that over one third of patients with epilepsy cannot achieve seizure control with available AEDs provides a powerful stimulus to research aimed at identifying new therapies. There is also a need for novel AEDs and AED formulations with improved tolerability profiles, which could replace older medications even in patients with currently responsive forms of the disease (55).

Although the consolidation taking place in the pharmaceutical industry may not encourage investment in therapeutic areas not perceived as strategic, there is no shortage of incentives to develop new medications for epilepsy (56). Because of the limitations of current treatments, a novel AED with an improved efficacy and tolerability profile could easily capture a large segment of the market. Many AEDs are also efficacious in other indications, so that the potential use of a new AED may extend far beyond its prescriptions for epilepsy. Finally, the cost of bringing an AED to the market is considerably lower than in other therapeutic areas, and special programs such as the National Institutes of Health Anticonvulsant Drug Development Program are in place to foster drug development in epilepsy. Against this background, it is not surprising that a number of potentially promising AEDs, some of which are reportedly active through innovative mechanisms, are in various stages of clinical development (Table 2) and many more are being evaluated preclinically (57). In addition to new chemical entities, innovative formulations (57) and innovative routes of drug administration (58) are also being tested. Among the latter, important breakthroughs could emerge from progress made in the on-line computerized analysis of EEG data used to predict the occurrence of seizures with an anticipation of several minutes before their clinical onset (59). Eventually, this could lead to “on-demand” intermittent administration of AEDs through a computer-driven device, an approach for which proof-of-concept has already been provided in animal models (60). Other exciting advances could come from the development of compounds that can interfere with the process of epileptogenesis, thereby preventing the occurrence of epilepsy in high-risk patients or antagonizing the progression of the seizure disorder that may occur in certain situations (61). Treatments aimed at providing a cure, such as gene therapy, may also become available in the future (58).

Table 2. A list of potential AEDs in clinical development
  1. AED, antiepileptic drug.

Carabersat (GlaxoSmithKline)
Conantokin-G (Cognetix, Inc.)
Pregabalin (Pfizer)
Retigabine (Viatris)
Rufinamide (Novartis)
RWJ 333 369 (Ortho-McNeil)
Safinamide (Newron)
SPM 927 (Schwarz Pharma)
SPD 421 (Shire)
UCB 34 714 (UCB Pharma)
Talampanel (Ivax)
Valrocemide (Teva)

While it is difficult to predict the extent to which each of these strategies will eventually affect the management of epilepsy, it is clear that exciting times are ahead of us.