The Long-Term Therapeutic Management of Epilepsy

Classification

The first step in effective management of epilepsy is correctly classifying and diagnosing the condition. The 1981 classification scheme [3], revised by the International League Against Epilepsy in 1989 to include epilepsy syndromes [4], is widely accepted. A simplified version is presented in Table 1. Epilepsy can be classified in terms of epilepsy syndromes as well as seizure types. Seizure type is determined by electroencephalographic patterns observed when the seizure is occurring and by clinical seizure characteristics. Examples include absence seizures (previously known as petit mal), simple partial seizures (previously known as focal motor or aura), complex partial seizures (previously known as “psychomotor” or temporal lobe), and generalized tonic-clonic seizures (also known as grand mal). Once seizure types are identified, this information determines the epilepsy syndrome in conjunction with information about cause, age at onset, and interictal electroencephalographic abnormalities.

Epilepsy can be basically classified as idiopathic (no apparent structural injury, probable genetic cause) or symptomatic (structural brain abnormality of known or unknown cause) (Table 1). Further differentiation can be made based on whether seizures begin in a focal region (partial epilepsy) or begin diffusely (generalized epilepsy). Idiopathic generalized epilepsy is the most common childhood epilepsy; absence, myoclonic, and generalized tonic-clonic seizures may be seen. In adults, symptomatic partial epilepsy is most common and may cause simple partial, complex partial, or generalized tonic-clonic seizures.

Some seizure types are specific to certain epilepsy syndromes, whereas others can be seen in various syndromes. It is often easy to determine the seizure type and the epilepsy syndrome on the basis of patient history and a routine electroencephalogram. For example, if a patient develops epilepsy after a head injury, the patient may describe seizures consisting of a warning of a peculiar smell followed by staring, lip smacking, and fumbling. The electroencephalogram shows a focal interictal spike. This evidence overwhelmingly points to partial epilepsy, manifested by simple partial seizures evolving to complex partial seizures. However, a substantial proportion of patients with epilepsy have confusing or conflicting information, making seizure and syndromic classification more difficult. For example, a patient may have no clear cause for epilepsy, may have seizures that are exclusively grand mal without warning, and may present with a normal routine electroencephalogram. Grand mal seizures may be the result of an idiopathic (genetic) generalized epilepsy or may occur in partial epilepsy if a focal onset is followed by rapid secondary generalization. In such a patient, every effort must be made to obtain further electrographic and historical information until the patient can be appropriately classified. Classification helps in selecting appropriate treatment, and in some circumstances, also determines prognosis and appropriate workup. Idiopathic epilepsy syndromes need little or no evaluation with imaging studies. In contrast, partial epilepsy is the result of a focal brain abnormality, and presentation of this syndrome in adults may warrant a search for underlying neoplasms or for other treatable diseases [5].

The choice of the initial anticonvulsant regimen differs, depending on the epilepsy syndrome classification. Table 2 indicates drugs of choice for the most common epilepsy syndromes and seizure types. When accurately diagnosed, seizures may be controlled by fewer or decreased concentrations of medication, with the potential advantages of fewer side effects, improved compliance, and enhanced well-being [6].

Selection of an Antiepileptic Drug

Once it has been determined that recurring seizures are not the result of a correctable problem, antiepileptic drug therapy should be initiated. Because all of the available antiepileptic drugs have different side effects, the types and severity of these effects must be considered when selecting an antiepileptic drug for an individual patient [7]. This decision must be made with the understanding of the long-term consequences because most patients receive medication for at least 2 years, and many continue for their lifetime. Unfortunately, with the currently available antiepileptic drugs, side effects are more the rule than the exception. Some patients tolerate medication without difficulty, but many have at least a mild alteration in some aspect of functioning. Clearly, life-threatening side effects are unacceptable for all patients, but less severe effects may be acceptable to one patient and not to another.

Because side effects differ somewhat from one drug to the next, the patient and that patient's specific situation in life must be kept in the physician's mind as initial therapy and subsequent choices are made. The physician and patient must become partners in the search for the regimen that produces the best quality of life.

Conventional Antiepileptic Drugs: Long-Term Considerations

The following discussion of toxicity is not meant to be encyclopedic. A specific discussion of antiepileptic drug toxicities can be found in many textbooks [8, 9]. Rather, this section focuses on long-term side effects and on issues that might affect long-term consequences of choosing one antiepileptic drug instead of another. Table 3 contains information about therapeutic ranges, half-lives, and pharmacokinetic properties for the drugs listed below.

Phenytoin

Although phenytoin (Dilantin;^Parke-Davis, Morris Plains, New Jersey) is primarily indicated for the management of tonic-clonic and partial seizures, it is also among the drugs of choice for treatment of most epileptic seizures, with the exception of absence and myoclonic seizures [10].

Within normal therapeutic ranges (5 to 20 µg/dL in most laboratories), phenytoin is usually well tolerated but may cause lethargy, abnormal movements, mental confusion, and cognitive changes [11]. At toxic concentrations that may be necessary to produce seizure control, phenytoin can cause nystagmus, ataxia, dysarthria, and encephalopathy [11]. Long-term phenytoin therapy can cause gingival hyperplasia, acne, hirsutism [12], and hypertrophy of subcutaneous facial tissue, resulting in an appearance known as “dilantin facies” [13]. These cosmetic effects may upset young women, and the potential consequences should be considered before beginning therapy. Scrupulous dental care can help prevent gingival hypertrophy, and visits to the dentist every 3 months are advised. Another long-term consequence of phenytoin therapy is osteoporosis [14]. If osteoporosis is diagnosed, vitamin D therapy may be necessary in conjunction with phenytoin, particularly for women [15]. Occasionally, a peripheral neuropathy can occur with chronic phenytoin use [12]. Cerebellar atrophy may also occur after decades of phenytoin therapy [16, 17]. The effects of phenytoin have been questioned, because cerebellar atrophy may also occur as a result of chronic frequent seizures. Nonetheless, if a patient manifests a cerebellar syndrome clinically, and a computed tomographic scan shows atrophy, switching to another agent should be seriously considered.

Like all anticonvulsants, phenytoin has teratogenic potential. Approximately 30% of infants exposed to phenytoin in utero develop mild craniofacial and digital abnormalities [12, 18]. In addition, from 1% to 11% of children exposed in utero to phenytoin develop a more severe group of anomalies, which may include congenital heart defects, cleft lip and palate, or microcephaly or mental subnormality or both. These abnormalities are known as the fetal anticonvulsant syndrome [19]. They are not unique to phenytoin and may be seen with all the commonly used antiepileptic drugs. Risks to the fetus can be kept to a minimum by using the lowest dosage of monotherapy necessary to control seizures, if possible, and by initiating folate therapy as soon as possible, preferably when the patient is contemplating pregnancy. Another disadvantage of phenytoin as long-term therapy is that it may produce drug-drug interactions. These may occur as a result of alterations in absorption or drug metabolism (most commonly due to interference with the hepatic cytochrome P-450 system). Some examples of clinically significant interactions include the ability of dicumarol, chloramphenicol, sulfonamides, isoniazid, disulfiram, and cimetidine to produce competitive or noncompetitive inhibition of hepatic metabolism and to increase serum phenytoin levels, whereas rifampin and antacids may decrease them [20-27]. Phenytoin, by inducing hepatic enzymes, may increase clearance of oral contraceptives, quinidine, chloramphenicol, and dicumarol [21, 27-29]. Salicylates may alter protein binding, causing an increase in free (active) phenytoin without altering the total measured serum level [30]. Other antiepileptic drugs, when coadministered with phenytoin, may produce either an increase or decrease in phenytoin level, which may be unpredictable [27].

Phenytoin has some advantages when considering long-term therapy. The first is that it is less expensive than many of the alternatives, which is important for patients who are paying for their own medications. Second, because of phenytoin's relatively long half-life Table 3, it can be given only twice a day and in some cases, once a day [12]. Studies have indicated that compliance improves considerably in patients who take medications less frequently [31], as discussed in a later section.

Carbamazepine

Carbamazepine (Tegretol;^Ciba-Geigy Corporation, Summit, New Jersey) is a first-line agent for simple partial, complex partial, and generalized tonic-clonic seizures [10, 11]. It may be contraindicated in patients with absence and myoclonic seizures [32]. Until recently, carbamazepine was thought to have less effect on cognition than phenytoin [33, 34]; however, this may not be true [35, 36].

Most adverse effects are related to serum concentration and include ataxia, drowsiness, vertigo, diplopia, and blurred vision [37]. Dividing daily dosages can frequently minimize these troublesome side effects [11]. Other adverse effects include nausea, vomiting, diarrhea, and constipation [37]. A common hematologic effect is a decrease in neutrophils, which does not cause immunosuppression [11]. In 2% of patients, severe leukopenia (with leukocyte counts < 2.5 × 10⁹) necessitates discontinuation of therapy [10].

The most serious side effect is bone marrow suppression with aplastic anemia, although recent reports indicate that this side effect is extremely rare [38]. Thus, carbamazepine is contraindicated for patients with a history of bone marrow depression [10]. Because carbamazepine was not approved for use until 1974, less is known about its long-term side effects [39]. Its use has not been associated with neuropathy, cosmetic changes, or cerebellar atrophy; osteoporosis is rare [40]. This may make it a better choice for long-term use. However, the short half-life Table 3, the requirement for dosing three to four times each day, and the expense when compared with phenytoin and phenobarbital are relative drawbacks. Similar to the other major antiepileptic agents, carbamazepine has been associated with an increased risk for malformations in the offspring of mothers taking the drug [41], including an increased risk for spina bifida [42].

Drug-drug interactions may complicate carbamazepine use. Carbamazepine induces hepatic enzyme metabolism. Therefore, clearance of other drugs, including theophylline, oral contraceptives, warfarin, and haloperidol, may increase. This also affects other antiepileptic drugs when given concurrently with carbamazepine, at times causing difficulty in attaining therapeutic levels [43]. In turn, carbamazepine clearance increases when given with these drugs. Inhibition of hepatic metabolism by erythromycin, isoniazid, propoxyphene, diltiazem, verapamil, and cimetidine may cause an increase in carbamazepine levels, frequently precipitating overt toxicity [44-49].

Phenobarbital

The barbiturate phenobarbital has been used since 1912, principally for controlling tonic-clonic and partial seizures [50]. It is widely used in neonates [51, 52]. The main adverse effects are sedation in adults and hyperactivity in children [53]. Other reported adverse reactions include memory loss, irritability, depression, and hematologic abnormalities (fetal vitamin K depletion, osteopenia, and megaloblastic anemia) [7, 54, 55]. In a comparative study, patients were less able to tolerate the side effects and were more likely to discontinue use of phenobarbital than they were to discontinue use of carbamazepine or phenytoin [34]. With long-term treatment, tolerance may develop to the sedative effect of phenobarbital [55]. When phenobarbital is administered on a long-term basis, connective tissue abnormalities may occur, including Dupuytren contractures, frozen shoulder, and general pain [56]. Appearance of these types of problems warrants consideration of alternate drug therapy. Despite the earlier feeling that the drug was safe in pregnancy, phenobarbital has been linked to teratogenicity. Compared with the rest of the population, almost a doubled risk exists for malformation [55]. The most commonly occurring teratogenic changes are cardiac anomalies and cleft palate [55].

Few problems exist related to absorption or protein binding of phenobarbital. Several drugs have the capability of inhibiting phenobarbital metabolism, including dicumarol, methylphenidate, propoxyphene, and most strikingly, valproic acid [55]. Similar to carbamazepine and phenytoin, phenobarbital induces hepatic enzymes and causes more rapid clearance of some drugs (see those listed for carbamazepine) [57].

Despite the considerable disadvantages of phenobarbital as long-term therapy, some patients tolerate it very well. For such patients, it may be an effective antiepileptic drug. Its major advantage is that it has, by far, the longest half-life of all the standard antiepileptic drugs, permitting once-a-day dosing [58] (Table 3). For patients who have a hard time remembering to take their medication or for those with erratic schedules, this can be a substantial advantage. A patient may take a dose 12 or even 24 hours too late and may still have a minimal drop in blood level. In addition, phenobarbital is the most inexpensive medication, costing only pennies a day compared with dollars a day for some other drugs.

Primidone

Once considered the drug of choice for complex partial seizures, primidone (Mysoline; Wyeth-Ayerst Laboratories, Philadelphia, Pennsylvania) is now used less frequently [59]. One study found it to be less effective and more toxic than carbamazepine, phenytoin, or phenobarbital [60]. Also, primidone is metabolized to phenobarbital, among other things, and causes all the problems seen with phenobarbital use, as described above [61].

Valproic Acid

Valproic acid (Depakene and Depakote;_Abbott Laboratories, North Chicago, Illinois) is one of the first-line agents for primary generalized epilepsies and absence seizures [62]. It achieves complete control in approximately 80% of patients with these seizure types, including those whose seizures are not controlled with other antiepileptic drugs [63]. It also is effective for generalized tonic-clonic convulsions seen in partial epilepsy; in a recent study [64], valproic acid was found to be less effective than carbamazepine for complex partial seizures. Unlike some of the other antiepileptic drugs, at therapeutic levels, valproic acid (see Table 3) has little or no effect on cognition [63, 65]. However, supratherapeutic levels that may produce cognitive disturbance and tremor are frequently necessary to control seizures [66].

Skin rashes are less common with valproic acid than with other antiepileptic drugs [7, 67]. The main side effects are drowsiness, nausea, vomiting, other gastrointestinal disturbances, tremor, weight gain, and transient hair loss [68]. The most serious effects are fatal hepatotoxicity, which is rare [69, 70], and pancreatitis [71]. Spina bifida is seen in up to 2% of children exposed to valproic acid before birth [72]. There is also a high incidence of other anomalies. These make valproic acid a poor choice for women contemplating pregnancy.

Valproate is metabolized almost exclusively in the liver. Concurrent use with other enzyme-inducing antiepileptic drugs may prevent attainment of adequate levels [73]. One clinically important interaction is seen between phenytoin and valproic acid, which has the capacity to displace phenytoin from its protein-binding sites [74]. This may cause emergence of phenytoin toxicity in the absence of an increased plasma level. Valproic acid is also highly protein bound and may be displaced. Clinically, however, this is rarely a problem [11].

Because valproic acid has been approved for use for only 15 years, long-term side effects are not known. Valproic acid may be the only effective agent for some seizure types [75-77]. However, it is expensive, and some patients are concerned about the considerable weight gain that can be associated with its use, which may necessitate discontinuation of valproic acid therapy [78]. Others are intolerant of tremor, which may be severe in some patients [79].

Ethosuximide

Ethosuximide (Zarontin, Parke-Davis) is an antiepileptic agent with a very narrow therapeutic indication. It is only effective as a first-line agent for absence seizures occurring in isolation, a condition seen primarily in childhood [80]. Occasionally, in primary generalized epilepsy with seizure types other than absence where valproate is ineffective, the addition of ethosuximide improves seizure control [80]. It is in this setting that most adult use of ethosuximide occurs. The most common side effects noted with ethosuximide use include nausea and abdominal discomfort, drowsiness, anorexia, and headache [81]. In rare cases, behavioral changes may be seen, including psychosis [82]. Teratogenicity has not been associated with ethosuximide [83]. Drug-drug interactions are also minimal. Ethosuximide is well tolerated during long-term use. However, as noted, its limited utility for adult epilepsy syndromes makes long-term use uncommon.

Benzodiazepines

Various benzodiazepines may be used for the management of epilepsy. Clonazepam (Klonopin; Roche Laboratories, Nutley, New Jersey) may be used as adjunctive therapy for seizures associated with the generalized epilepsies but is less effective for partial epilepsy [84]. Clorazepate (Tranxene, Abbott Laboratories) may be used for both generalized and partial epilepsies [85, 86].

Benzodiazepines are not good choices for long-term therapy and should not be used as first-line agents. Patients frequently develop tolerance to the therapeutic effects of the benzodiazepines. Seizures may initially be decreased, but after weeks to months, seizures recur, necessitating dosage increases to regain control. Frequent dosage increases exacerbate the most common dose-related side effects of these agents, such as drowsiness, ataxia, and behavioral problems [84, 86]. In addition, benzodiazepine withdrawal may precipitate status epilepticus [87].

Monotherapy Compared with Polytherapy

Single antiepileptic drug therapy, or monotherapy, is now the preferred method of treatment for epilepsy. Polytherapy increases the risk for drug interactions, teratogenicity, and adverse effects; makes monitoring more difficult; and results in decreased patient compliance [6, 31, 34, 53, 60, 65, 88, 89].

Studies have indicated that seizure control can be obtained with optimal monotherapy in about two thirds of patients [7]. In fact, one large study showed that the addition of a second antiepileptic drug achieved total seizure control in only 11% of patients, whereas toxicity developed in 90% of patients [90]. Conversely, one study showed that converting the therapeutic regimen from combined therapy to monotherapy produced no increase in seizures in 83% of patients and produced a decreased frequency of seizures in 36% [53]. As these studies show, the benefits of polytherapy are not proportional to the added risks for toxicity.

Physicians managing patients on combined therapy must be aware of the drug interactions associated with concurrent use of antiepileptic drugs in order to make appropriate therapeutic decisions. Although certain interactions may be minor, others can be extremely dangerous [31]. Even when serum levels of the antiepileptic drugs are known, identifying which drug is responsible for a toxic reaction may not be possible. Drug interactions can occur even when neither drug is present in the blood in unacceptable ranges [31].

Carbamazepine, phenytoin, phenobarbital, and primidone have enzyme-induction properties [31]. This tends to decrease the half-lives of concurrently administered antiepileptic drugs compared with the half-lives achieved with monotherapy and may lead to subtherapeutic levels at the end of the dosing interval, unless each antiepileptic drug is administered more frequently or in higher doses. In addition, combining antiepileptic drugs that stimulate metabolism of endogenous substances may cause inactivation of vitamin D, thereby placing patients at greater risk for rickets and osteomalacia, as well as for increased metabolism of sex hormones, which increases the probability of acne, hirsutism, and decreased libido [31].

Recent evidence suggests that switching patients on polytherapy to monotherapy decreases these adverse effects [53, 89]. In one study [90], patients who were converted to monotherapy showed a 55% improvement in cognitive function, particularly alertness, drive, mood, concentration, and sociability. For patients whose seizures are resistant to monotherapy within therapeutic ranges, higher doses are often administered. Even at these doses, treatment with a single drug has been associated with equal or better seizure control and with lower toxicity than polytherapy [54, 88].

Therefore, the most important principle in long-term management of epilepsy is that therapy should always be initiated with a single antiepileptic drug [31, 91]. The drug should be initiated at a moderate-to-low dose and should be slowly increased while assessing therapeutic effect. The patient should be maintained on the lowest effective dose. If low dosages are not effective, the medication should be pushed to toxicity before it is abandoned, even if the resultant blood level is above the laboratory definition of “therapeutic.” If intolerable side effects develop before adequate seizure control is obtained, the patient should be switched to a second single drug (monotherapy). Only when all appropriate drugs in monotherapy have failed should combination therapy be tried. A more difficult decision may exist when a physician assumes care for a patient who is already receiving polytherapy. If the patient is not controlled or is having substantial side effects, the drug presumed to be most appropriate and least toxic for long-term therapy should be continued and the other drug(s) discontinued. The physician may then be able to increase the dosage of the primary drug, without causing side effects.

The long-term goal may not always be the complete eradication of seizures at any expense. Depending on the social situation, the patient's temperament, and other factors, a patient may prefer to have medication decreased, have side effects diminished, and have a few seizures. This is particularly true if the resultant seizures are nocturnal or are simple partial seizures, which may not result in driving restrictions. Other patients may want to be seizure-free at any cost.

Monitoring of Drug Levels

One issue in the long-term care of patients with epilepsy is the frequency and use of obtaining plasma levels of antiepileptic drugs. Most laboratories provide a “therapeutic range,” the range within which most patients have a therapeutic effect without having symptoms of toxicity (Table 3). These ranges are only the most basic guideline. Patients may become seizure-free at so-called subtherapeutic levels or may have toxic reactions at levels within the therapeutic range. This is particularly true for patients on polytherapy. In a recent study [92] of side effects, researchers reported no substantial relation between dose-dependent side effects and plasma levels. Some patients complained of these side effects before their plasma levels were in the therapeutic range, whereas others tolerated much higher doses and levels [92]. Again, patients may be maintained safely on medication doses producing blood levels above the therapeutic range, if they are free of side effects and if lower levels do not control seizures. While medications are being titrated, levels may be obtained to assess metabolism of the drug in that individual patient, because the same dosage does not necessarily produce the same level in two patients of the same height and weight.

Monitoring requirements differ according to particular antiepileptic drug regimens. When phenytoin is administered in higher doses, first-order elimination kinetics are replaced by zero-order kinetics. When this happens, slight changes in dose can result in wide swings in plasma concentrations [93]. Monitoring plasma levels can, therefore, help to determine the optimal phenytoin dosage for a particular patient [82].

Carbamazepine has a short half-life, particularly in children, and plasma levels can fluctuate widely from below target range to toxic levels within a single day [93]. Obtaining a profile of carbamazepine levels during a 24-hour period can show the peaks and troughs and can lead to dosage changes. Valproic acid has a relatively short half-life, with plasma levels fluctuating during the day. However, this fluctuation does not appear to be linked to the efficacy of the drug [93].

Serial blood levels can be used to assess compliance. Whether assessing compliance or obtaining levels for other reasons, levels should always be drawn at the same time of day and preferably at trough, just before a dose [94]. This is particularly true for drugs with short half-lives. Drawing standard levels of the parent compound may not always be sufficient to assess the efficacy and the toxicity of antiepileptic drugs. Some drugs, most notably carbamazepine and primidone, have metabolites that contribute to toxicity and sometimes to efficacy [94]. Other drugs, most notably phenytoin and valproic acid, are highly protein bound. Only the free drug is active; the percentage of protein binding may change with nutritional state, concurrent medication administration, pregnancy, and disease [94]. Free levels should be drawn in these conditions or if excessive side effects are noted at therapeutic levels. Once a patient has been stabilized on a regimen, biannual or annual levels may be sufficient for long-term follow-up. Because metabolism may change over time, these levels allow comparison to a level obtained at the time of a breakthrough seizure. However, subtherapeutic levels should not lead to medication changes if the patient is seizure- and symptom-free. Additional levels may be necessary if new side effects occur.

Discontinuation of Antiepileptic Drug Therapy

Patients who have been seizure-free for long periods may be candidates for discontinuation of medication. Studies [95] indicate that 40% to 70% of patients who have remained seizure-free for 2 to 3 years or more may have their medication discontinued, without return of seizures. Although it is impossible to predict which patients can be withdrawn successfully and which will have a recurrence, some factors may indicate a higher or lower likelihood of relapse. Although controversy exists about this issue, a normal electroencephalogram, a shorter duration of epilepsy, presence of absence seizures only, and younger age when seizures came under control are considered good prognostic signs [95]. Although not mandated by law in most states, it is advisable to tell the patient to refrain from driving during withdrawal and for the following few months, the time when recurrence is most likely [95].

The patient must be the one, ultimately, to decide whether the risk for seizure recurrence is worth the benefit of discontinuing medication. This decision depends on whether the patient is having chronic side effects and on the consequences of a seizure on the patient's life. Some patients may decide to remain on medication indefinitely.

Medication Failure

Misdiagnosis

At some point, after one or several trials of monotherapy have failed to adequately control seizures with acceptable side effects, it is reasonable to reevaluate the patient and to determine whether treatment is failing because the patient has not been accurately diagnosed. The patient may have been given the wrong syndromic classification or may not have epilepsy at all. In some studies [96-98], 7% to 25% of apparently intractable patients actually were having nonepileptic (pseudoepileptic, psychogenic) seizures, either alone or in combination with epileptic seizures.

Ideally, for appropriate diagnosis and classification, all intractable patients should have an abnormality visible on an electroencephalogram. However, the presence of an epileptiform electroencephalographic abnormality is not synonymous with the diagnosis of epilepsy. Patients with frequent seizures may have completely normal electroencephalograms between events [99, 100]. Conversely, patients with nonepileptic seizures may have abnormal or even epileptiform electroencephalograms [101]. If a routine electroencephalogram has failed to show an abnormality, a sleep-deprived electroencephalogram or an outpatient ambulatory electroencephalogram may be helpful. If all of these have been done, and the diagnosis is still in doubt, inpatient or outpatient video-electroencephalographic monitoring may be necessary [102]. Video-electroencephalographic monitoring involves the use of simultaneous closed-circuit television and electroencephalographic recordings to show the sequence of behavioral and electrical events that occur during seizures. This technologic development is useful in capturing interictal electroencephalographic abnormalities, diagnosing unusual cases of epilepsy, and distinguishing epileptic from nonepileptic seizures [103].

Physicians should compare electroencephalographic and monitoring results with the International League Against Epilepsy classification scheme (see Table 1) to determine whether the proper diagnosis has been assigned and should then determine whether the current drug regimen is the appropriate monotherapy for this seizure type.

Noncompliance

A critical issue in long-term management of epilepsy is noncompliance with prescribed antiepileptic drugs, and it may be the single most common reason for antiepileptic drug failure [31, 104, 105]. A large percentage of adult patients (40% to 60%) and pediatric patients (25% to 75%) do not comply with their prescribed regimens [11]. In a recent study [106], overall noncompliance was estimated at 63%. Research [107] shows that physicians overestimate their patients' compliance by approximately 50%, and even when they become more familiar with their patients, their estimates of compliance do not improve. In patients with chronic disorders, such as epilepsy, noncompliance is especially problematic. Before replacing an antiepileptic drug that appears unable to control seizures, it is crucial to establish whether the lack of seizure control is due to drug failure or to the patient's failure to take prescribed dosages.

No association exists between noncompliance and factors such as age, sex, level of education, and duration of the disease [106]. However, an association has been established between noncompliance and the number of concurrent medications prescribed. Noncompliance is clearly higher in patients requiring combined antiepileptic drug medication than in patients receiving monotherapy [106, 107]. Noncompliance also increases when patients misunderstand instructions for using prescribed medications [108]. In a recent study [106] of phenytoin and theophylline compliance, approximately 50% of patients with chronic disorders did not realize how crucial it was to take their prescribed medications continuously. Because a certain concentration of serum antiepileptic drug must be maintained, skipping a dose or taking several doses simultaneously can have a dramatic effect on seizure control.

Many patients or their physicians may not be certain whether patients are adhering to their antiepileptic drug regimen. For those patients, the Medication Event Monitoring System (MEMS) can measure compliance on a daily and an hourly basis and is minimally intrusive. The Medication Event Monitoring System involves use of a special 15- or 30-dram pill bottle, fitted with caps that contain a microprocessor that can record as many as 1000 openings. The Medication Event Monitoring System records the patient's adherence to the prescribed dosing schedule [108, 109]. Pill counting can also determine a patient's adherence to a prescribed regimen. According to this method, the number of leftover pills or days between prescription refills is counted. This method, however, has drawbacks. Unused bottles may be mislaid or deliberately not returned, extra doses could be consumed to make up for missed doses, and no record of hourly or daily adherence to the prescribed dosing schedule is provided.

If a patient has not adhered to a prescribed antiepileptic drug regimen, the physician or support personnel should provide counseling and reinforcement [110]. It is also helpful to suggest memory aids, such as linking doses to events in the patient's daily schedule (meals, brushing teeth) or purchasing alarmed watches or pill cases. The physician may also convert to a therapy that requires less frequent daily dosages if compliance remains a problem.

Other Treatment Options

When all forms of antiepileptic drug monotherapy prove unsuccessful, physicians should prescribe one or two simple drug combinations. If this combined therapy proves unsuccessful, the antiepileptic drug regimen that provided the best seizure control and the fewest number of side effects should be continued. Simultaneous use of more than two drugs is rarely indicated.

Other options, including second-line drugs, experimental drugs, or surgery should be considered when standard drugs have failed. Again, patients must decide whether these options are acceptable to them. All or most of these patients will benefit from referral to a comprehensive epilepsy center. Such centers can be located by contacting the National Association of Epilepsy Centers, Minneapolis, Minnesota.

Surgery

Of those 25% of patients with epilepsy who have intractable seizure disorders [111], between 12% and 25% are candidates for epilepsy surgery [112]. Good surgical candidates should meet the following criteria: medically intractable epilepsy, clinically significant disease that interferes with patient's lifestyle, and medical and psychological stability [113, 114]. The evaluation process is long and arduous. However, the appropriately selected surgical patient has a good chance of becoming seizure-free and remaining so.

Nontraditional Agents

When conventional antiepileptic drugs fail, it is sometimes appropriate to attempt therapy with “second-line” agents. These agents are not the agents of choice because they have higher toxicity or lower efficacy, or both, compared with standard antiepileptic drugs. Nevertheless, individual patients may respond favorably. Phenacemide (Phenurone, Abbott Laboratories) is recommended for treatment of complex partial seizures when conventional antiepileptic drugs are unsuccessful and surgery is not an option [115]. Side effects are dose related. Use of this drug may result in ataxia, drowsiness, and attention deficits. Rare but serious adverse effects include liver failure and aplastic anemia, sometimes resulting in death [116].

Mephenytoin (Mesantoin; Sandoz Pharmaceuticals Corp., East Hanover, New Jersey), an antiepileptic drug that fell out of favor because of dangerous idiosyncratic adverse effects (agranulocytosis, aplastic anemia, and more rarely, exfoliative or necrolytic dermatitides), is now the focus of new attention [117]. Like phenytoin, this agent is used to treat all seizures of focal origin and all convulsive seizures and has an efficacy profile similar to that of phenytoin. It is primarily used in patients who have responded to phenytoin but who have had undesirable side effects. Mephenytoin is less likely to produce ataxia, mental dulling, collagen and connective tissue overgrowth, and peripheral neuropathy compared with phenytoin but is more likely to produce drowsiness [117]. All patients given mephenytoin should have a complete blood count monthly during the first year after initiation of therapy in order to screen for dangerous idiosyncratic hematologic effects [117]. Methsuximide (Celontin, Parke-Davis) therapy is recommended for patients with complex partial seizures refractory to first-line antiepileptic drug therapy. Common dose-related adverse effects include gastrointestinal disturbances, lethargy, somnolence, fatigue, and headaches. Hiccups, irritability, ataxia, blurred vision or diplopia, inattention, dysarthria, and psychic changes are also found [118]. Transient leukopenia has also been reported [119]. Methsuximide interacts with other antiepileptic drugs, suggesting that close monitoring of serum levels and dosage adjustments of concurrent antiepileptic drugs may be indicated [119].

Emerging Antiepileptic Drugs

Although epilepsy is one of the most common neurologic disorders, the clinician must still rely on antiepileptic drugs developed decades ago. Many patients with epilepsy must accept bothersome adverse effects or diminished quality of life to achieve complete seizure control. No antiepileptic drug exists that is free from teratogenic potential or confusing drug-drug interactions. Clearly, new antiepileptic drugs are urgently needed that are effective but produce fewer adverse effects.

Many investigational drugs are currently available in the context of controlled clinical trials. Patients may be referred to epilepsy centers where these trials are conducted. The patient must return for frequent visits and must know of the possibility of receiving placebo in the context of a blinded trial. In return, however, the company that manufactures the drug usually pays for all medical care related to the trial, and all patients can receive the drug in an open manner once the trial is complete.

Several of these new agents are nearing or have received approval by the Food and Drug Administration. Ultimately, some may be as effective or more effective than currently used agents and may have fewer side effects. In that case, it may be necessary to reevaluate the order in which drugs are tried, with newer agents used before more traditional choices. Three agents that have been or will be approved in the near future are briefly described below.

Felbamate (Felbatol; Carter-Wallace, Cranbury, New Jersey) is a novel compound, structurally similar to meprobamate, but it does not have the sedating properties of this compound [120, 121]. In fact, it appears to be well tolerated. Potential side effects include nausea, headache, weight loss, and insomnia [120, 121]. No teratogenicity has been found in animals [121]. Felbamate has a broad spectrum of anticonvulsant action [120]; it is efficacious for partial epilepsy [122, 123]. A recent trial [124] also showed efficacy in the Lennox-Gastaut syndrome, a type of epilepsy associated with mental retardation and multiple seizure types, which is usually refractory to medical therapy. Felbamate inhibits hepatic metabolism and produces drug-drug interactions. Phenytoin and valproate levels increase when felbamate is given concurrently. Total carbamazepine levels decrease, but levels of carbamazepine epoxide (an active metabolite) increase. These interactions should be considered as the potential cause of side effects when felbamate is used in polytherapy [121].

Gabapentin (Neurontin, Parke-Davis) is a γ-aminobutyric acid (GABA) analog absorbed across the blood-brain barrier. However, it does not interact with γ-aminobutyric acid receptors in the central nervous system, and its mechanism of action remains undefined [125]. In patients with particularly resistant partial epilepsy, gabapentin decreased the incidence of seizures significantly, with few antiepileptic drug-drug interactions detected [126].

Lamotrigine (Lamictal; Burroughs Wellcome Company, Research Triangle Park, North Carolina) is a novel antiepileptic drug that seems to act by inhibiting the release of glutamate, aspartate, and, to a lesser degree, γ-aminobutyric acid. In clinical trials, lamotrigine has shown efficacy in patients with partial seizures (particularly those that are secondarily generalized) as well as primary generalized seizures, and it has a favorable safety profile [127-129]. Several interactions with other antiepileptic drugs have been noted [128, 130].

Social, Vocational, and Psychological Assistance

Epilepsy is a condition that affects all aspects of a patient's life, including personal relationships, employment, and social functioning. Part of the physician's duty in providing long-term care is to make services that address these areas available to the patient. Many of these services are provided by local branches of the Epilepsy Foundation of America. These organizations, as well as many epilepsy centers, provide epilepsy support groups. Psychiatric support or psychological counseling may also be necessary. Epilepsy has been linked to an increased incidence of some psychiatric disorders, particularly depression [131].

Although much effort may be expended by the physician, it may not always be possible to improve the patient's epilepsy. The constant search for a new therapy is important and allows patients to maintain hope. However, anticipating better therapy in the future may prevent patients from accepting their condition and from adjusting accordingly; these patients may require psychological counseling.

Long-term management of epilepsy requires good communication between the physician and the patient. Appropriate therapeutic strategies can only be determined by assessing the patient's needs as a whole. Most medical and surgical options that are currently available have some drawbacks. Despite this, most patients, if appropriately managed, can achieve seizure control with an acceptable level of side effects. We should hope that new advances in surgical procedures and in drug therapy will offer more acceptable options in the future.