A nurse is caring for a patient who has tonic clonic seizures and is prescribed phenytoin therapy

Introduction
Epilepsy is a common neurological disorder. Over 2% of the population will have a non-febrile seizure at some time in their life, and over 1% of the population have unprovoked, recurrent seizures (epilepsy). Thus, seizures present common management problems in medical practice. Although anticonvulsants are used mainly for epilepsy, they have other uses e.g. in the management of trigeminal neuralgia, neuropathic pain syndromes and mood disorders.1

Classification
Two major questions determine the decision to treat a patient and the choice of therapy: is this epilepsy, and if so, what is the classification of the seizures and the epilepsy? Not every paroxysmal event is a seizure, and misidentification of other conditions as epilepsy can lead to unnecessary and potentially harmful treatment. Accurate diagnosis rests upon the clinical details, in association with electro-encephalographic and laboratory studies. Correct classification is essential when managing a patient with epilepsy. It influences the decision about whether to treat, with what drug, and for how long.

Seizures and epilepsy are currently classified according to a scheme developed by the International League Against Epilepsy (ILAE). This recognises two broad categories of seizures: those arising from one hemisphere (partial or focal), and those which arise from both hemispheres simultaneously, the latter being subclassified by the presence of patterns of convulsive movements. Seizures are the clinical symptoms and signs of epilepsy. The seizure types, together with the age of onset, family history, and associated clinical and laboratory findings help determine the type of epilepsy. This diagnosis is important with respect to drug therapy and also carries prognostic implications. A simplified version of the scheme is shown in Table 1, and some old terminology is included for reference.

Selecting an antiepileptic drug
(Tables 1 and 2)

When the decision has been made to treat a patient who has epilepsy, there are two main considerations:

• efficacy - can the seizures be controlled?
• toxicity - can adverse effects of the treatment be minimised or prevented?

Partial and secondarily generalised seizures respond equally well to carbamazepine, phenytoin, phenobarbitone and methylphenobarbitone, although in any patient one drug may be more effective. These drugs differ substantially in their adverse effects and this, coupled with pharmacokinetic properties, helps determine the choice of drug for particular patients. Phenytoin, with its relatively long half-life, can be given once or twice daily. For some patients this may be preferable to carbamazepine, which must be taken more frequently. Concern about the occasional undesirable cosmetic adverse effects of phenytoin makes carbamazepine the drug of choice for other patients. Both phenobarbitone and methylphenobarbitone have a high incidence of sedation when treatment commences. They should not be considered as first-line drugs in adults with epilepsy.

Primary generalised seizures respond best to sodium valproate which can be used effectively as monotherapy in up to 80% of patients. Myoclonic seizures respond to benzodiazepines, especially to clonazepam. Ethosuximide is the drug of choice for absence seizures. It is as effective as sodium valproate and has fewer adverse effects.

Drug concentration monitoring
The management of epilepsy has been greatly influenced and enhanced by the introduction of rapid, and reproducible, drug assays. A close relationship between serum concentration and therapeutic and toxic effects exists with some antiepileptic drugs (e.g. phenytoin and carbamazepine), but is less clear cut with others (e.g. sodium valproate, phenobarbitone and ethosuximide). However, the 'therapeutic range' is only a guide to dosage adjustments. The clinical response to therapy, symptoms and signs of toxicity, administration history and sampling details must be considered when interpreting serum concentrations. Some patients have adverse effects at low or therapeutic concentrations, while others tolerate high levels without the development of adverse effects.

Table 1

Classification of seizures and drugs used to treat them (in the author's order of preference)
(see text). Some of the old terminology is included in brackets.

Table 2

The dose, half-life, 'therapeutic range' and commonly observed adverse effects of the most widely used antiepileptic drugs

There is also considerable inter individual variation in the relationship between the serum anticonvulsant concentration and seizure control. The interpretation of serum concentrations should also take into account a drug's pharmacokinetics. Sodium valproate has a wide therapeutic index, large fluctuations in its concentration-time profile and concentration-dependent protein binding.

Carbamazepine has a flatter concentration-time profile, a more clearly defined target range, undergoes auto-induction of metabolism and interacts with other drugs. Phenytoin has non-linear kinetics and a narrow therapeutic range. This is of considerable practical importance because small changes in the dose can have profound effects on the plasma drug concentration, which may result in toxicity.

Drug concentrations are especially useful in:

• monitoring therapy with phenytoin and carbamazepine
• assessing compliance with therapy
• patients with renal or hepatic disease (in whom both the free and total levels are useful)
• concomitant therapy for other medical illnesses
• the elderly
• pregnancy

The adverse effects of the antiepileptic drugs may be divided into dose dependent and allergic or idiosyncratic reactions which are unrelated to the serum concentrations. In addition, adverse effects may be related to the formation of a metabolite which has been proposed for carbamazepine 10, 11 -epoxide. Table 2 lists the commonly observed adverse effects of some antiepileptic drugs.

Starting therapy
Patients should be managed, where possible, with a single antiepileptic drug. With the exception of phenytoin (which can be introduced at a maintenance dose), most antiepileptic drugs should be introduced slowly to minimise adverse effects. In general, the interval between doses should be less than one third to one half the drug's half-life at steady state. The dose should not be increased until a steady state has been achieved (i.e. an interval of 5 times the drug half-life).

Carbamazepine should be introduced at low dosages (200 mg daily in adults and adolescents) as it may produce psychomotor impairment. The dose may be increased shortly after commencing therapy as auto-induction of metabolism and tolerance to the central nervous system adverse effects occur early. Children require higher dosages of carbamazepine on a mg/kg basis than adults.

Sodium valproate can be introduced at a dose of 500 mg daily. Higher starting doses may cause gastrointestinal adverse effects.

Phenytoin can take up to 4 weeks to reach steady state and should be commenced at a dosage of around 5 mg/kg/day. If urgent control is needed, a loading dose of 10-20 mg/kg can be administered, either orally or intravenously.

Ethosuximide does not require a loading dose, but should also be introduced slowly to minimise the potential gastrointestinal adverse effects.

Maintenance dosage
The maintenance dose of carbamazepine may need to be increased during the first few months of therapy because of the auto-induction of metabolism. This effect varies significantly between patients. The adjustment of the dose of sodium valproate should be made according to clinical observation rather than serum concentration measurement. This is because the concentration-response-toxicity relationship is unclear, the drug has a wide safety margin and there is non-linear protein binding at higher concentrations.

Optimisation of the phenytoin dosage is complicated by its non-linear pharmacokinetics. Small changes in the dose of phenytoin may move the serum concentration out of the narrow therapeutic range.

Ethosuximide dosage should be adjusted according to the clinical response.

Polytherapy
While at least 80% of patients can be adequately controlled on one drug, some may require a second drug to obtain maximal seizure suppression. Drug interactions can alter dose-concentration relationships and there is a significant increase in adverse, especially sedative, effects with polytherapy. It is difficult to predict the extent and, at times, the direction of the pharmacokinetic interactions. Therefore, drug concentrations should be monitored when a second drug is introduced and when a second or third drug is withdrawn. Little information is available on the correlation between plasma levels of anticonvulsants used in combination and the therapeutic efficacy or toxicity. Dosages have to be manipulated on the basis of clinical judgement and only one drug dose should be altered at a time. When managing patients who have incompletely controlled seizures by combination therapy, the prescriber should be careful that adverse effects do not become a greater handicap to the patient than the seizures themselves.

Several new antiepileptic drugs are in the process of clinical trials or about to become available for clinical use. Vigabatrin, an irreversible inhibitor of gamma-aminobutyric acid (GABA) transaminase, has been approved as add-on therapy for patients whose seizures are controlled incompletely by other drugs. The drug is particularly useful in subjects with partial seizures, in who seizure frequency may be halved. Readily absorbed from the gut, the drug is excreted mostly in the urine. The clinical efficacy of vigabatrin is not well correlated with serum concentrations and the dosage should be adjusted according to the clinical response and patient tolerance. An interaction with phenytoin may result in a fall in serum phenytoin concentration by up to 20%.

When seizures are refractory to antiepileptic drug treatment, the diagnosis should be reviewed. Surgical treatment is used increasingly in refractory cases.

(See also Dental implications and Teratology comments.)