|Home | About | Journals | Submit | Contact Us | Français|
Epilepsy is a chronic condition requiring long-term treatment with drugs that have intrinsic limitations. Antiepileptic drugs (AEDs) are effective in suppressing seizures but do not alter the disease process. They have a suboptimal tolerability profile and can be teratogenic. Second-generation compounds may be better tolerated but no more effective than traditional AEDs. In this light, as drug therapy is purely symptomatic, acute symptomatic seizures (i.e. seizures occurring in close temporal relationship with acute CNS insults) may require treatment only until recovery or stabilization of the injury. Treatment of the first unprovoked seizure may be considered in patients with abnormal EEG and imaging findings and in those in whom the relapse has severe social, emotional and personal implications. In these cases and in patients with epilepsy (i.e. repeated unprovoked seizures), drugs for partial seizures supported by class I regulatory trials or pragmatic trials are oxcarbazepine in children, carbamazepine or lamotrigine in adults, and lamotrigine or gabapentin in the elderly. Pragmatic trials support use of valproate for generalized seizures, except for women of childbearing age for whom the drug should be tailored to the individual patient. The lowest maintenance dose should be chosen, based on the efficacy and tolerability of the assigned drug. If the first monotherapy fails, the safety profile of a drug is important when opting for another monotherapy or for an add-on therapy. The epilepsy syndrome and the social, psychological and emotional profile of the patient all contribute to the individualization of treatment discontinuation after long-term seizure remission.
Epilepsy is a treatable clinical condition. The goal of treatment is the achievement of seizure freedom after treatment start, during treatment, and after drug withdrawal. An increasing number of drugs have been made available in the last few decades for the treatment of epilepsy with undisputed benefits in terms of seizure control. However, the use of the available antiepileptic drugs (AEDs) is subject to a number of limitations which must be considered when long-term treatment is initiated in a patient with epilepsy. First of all, AEDs are ‘symptomatic' (antiseizure and not antiepileptogenic) agents. Both first-generation and second-generation compounds are effective in suppressing seizures but they do not seem to alter the underlying disease process [Schachter, 2002]. Although an antiepileptogenic effect has been observed in experimental animals with some drugs [Pitkanen, 2002; Loscher et al. 1998], this effect has not been confirmed in humans (see below). Second, AEDs have a suboptimal tolerability profile [Zaccara et al. 2007; Walia et al. 2004] and can be teratogenetic [Perucca, 2005]. Third, second-generation AEDs have better pharmacokinetic profiles [Bialer, 2005] and are – to some extent – better tolerated but no more effective than traditional AEDs [Beghi, 2004]. Last, the drug of choice is dictated by several factors which may differ across patients. With these limitations in mind, the treatment of epilepsy has been assessed across the following stages: (1) treatment of acute symptomatic seizures, (2) treatment of the first unprovoked seizure, (3) treatment of newly diagnosed epilepsy, and (4) treatment discontinuation.
The decision to start antiepileptic treatment is made when the purported effects of the treatment are supposed to outweigh the risks. The benefits of treatment are measured in terms of impact on seizure recurrence which in turn is affected by the definition of seizures and epilepsy. Epilepsy is defined by the occurrence of two or more unprovoked seizures [Commission on Epidemiology and Prognosis, International League Against Epilepsy, 1993]. An unprovoked seizure is a seizure or a cluster of seizures occurring within 24 hours in a person older than 1month of age, occurring in the absence of precipitating factors. Unprovoked seizures may be single or recurrent. Although patients with single unprovoked seizures may have an enduring predisposition to epilepsy [Fisher et al. 2005], seizure recurrence can be observed only in about one-half of cases [Berg and Shinnar, 1991]. In contrast, after a second unprovoked seizure, the risk of a third seizure has been estimated as 73% and the risk of a fourth seizure as 76% [Hauser et al. 1998]. In contrast, acute symptomatic seizures are seizures occurring at the time of a systemic insult or in close temporal association with a documented brain insult [Commission on Epidemiology and Prognosis, International League Against Epilepsy, 1993]. Acute symptomatic seizures differ from unprovoked seizures in terms of mortality and recurrence [Hesdorffer et al. 2009] and can be considered risk factors for unprovoked seizures and epilepsy but not symptoms of epilepsy per se. In this context, the results of randomized clinical trials and meta-analyses consistently showed that in patients with differing CNS insults (febrile seizures, head trauma, infection, asphyxia, toxic agents, brain tumor, craniotomy) AEDs are effective in controlling acute symptomatic seizures but they fail to prevent unprovoked seizures and epilepsy [Temkin, 2001](Table 1).
In summary, AEDs are successful for the prevention of acute symptomatic seizures but apparently not for the prevention of unprovoked seizures and epilepsy. This supports the concept that the available drugs are purely symptomatic with no antiepileptogenic action. The treatment of acute symptomatic seizures depends on the duration of the period at risk (i.e. the interval between the onset of the CNS injury and the time of recovery or stabilization) which varies according to the underlying clinical condition [Beghi et al. 2009]. Acute symptomatic seizures may occur during the week following a stroke or a severe head trauma. In the presence of a cerebral infection, however, these seizures may persist for the entire period of activity of the infection (indicated by the persistence of symptoms, signs or laboratory findings). For these reasons, treatment in these cases must be individualized.
The cumulative probability of relapse of a first unprovoked seizure at 2years is 42% [Berg and Shinnar, 1991]. The two most consistent predictors of recurrence of a first seizure are an abnormal (particularly epileptiform) EEG and the presence of an underlying symptomatic cause or an abnormal neurological exam. The two factors may have an additive effect in that patients with both a symptomatic cause and an abnormal EEG appear to have an even higher risk than those with only one of those factors [Kim et al. 2006]. Other prognostic predictors include the occurrence of the seizure during sleep, seizure type (partial seizures are followed by a higher risk of relapse compared with primary generalized seizures), and family history of seizures and epilepsy [Berg and Shinnar, 1991]. Status epilepticus and a history of febrile seizures may be also associated with an increased risk of recurrence in individuals with symptomatic seizures [Berg, 2008]. These factors should be considered in the decision to treat a first unprovoked seizure. Other variables, however, include the presence of concurrent medical problems (requiring chronic treatments), patient compliance, legal implications (fitness to drive, insurance), and the patient's sociocultural background, emotional problems, and the effects of treatment on seizure relapse and long-term seizure remission. A number of randomized trials comparing AEDs with placebo or no treatment in patients with a first unprovoked seizure have consistently shown the efficacy of drugs in preventing seizure relapse but no effects of treatment on the achievement of prolonged seizure remission [Beghi, 2007](Table 2). The results are mostly driven by two large-scale randomized trials, the multicenter study from Italy, FIR.S.T [Musicco et al. 1997; First Seizure Trial Group, 1993] and the European-wide Multicenter Epilepsy and Single Seizure study or MESS, the latter including both first seizures and epilepsies for which the caring physician was uncertain about treatment [Marson et al. 2005].
In summary, indiscriminate treatment of the first unprovoked seizure with AEDs is not recommended [Beghi et al. 2006]. Treatment may be considered in patients in whom EEG and imaging data indicate an increased risk of relapse (presence of structural CNS and/or EEG abnormalities) and in those in whom the risks and the benefits of treatment are in favor of the latter, after consideration of the social, emotional and personal implications of seizure relapse and of treatment itself. There are situations which may indicate deferral of treatment (e.g. pregnancy) while others (e.g. patients performing potentially dangerous activities) may favor initiation of treatment. The patient should be involved in the decision process. Treatment modalities (choice of drug, drug dosages and duration of treatment) are the same as for the treatment of patients who had recurrent seizures (see below).
A recent guideline has been recently issued by experts from the International League Against Epilepsy (ILAE) who undertook a systematic review of the literature on the choice of monotherapy in patients with newly diagnosed epilepsy [Glauser et al. 2006]. The published reports were grouped in three categories (class I, II, and III) based on the robustness of the study design and the quality of data. Every report was tabulated with reference to the class of the report, the type of epilepsy (partial or generalized) and the patient's age (children, adults, and elderly). Accordingly, the therapeutic recommendations were expressed in three categories (A: definite efficacy; B: probable efficacy; C: possible efficacy). Based on type A or type B recommendation, the drug in question should be considered for initial monotherapy. Based on type C recommendation, the drug might be considered for initial monotherapy but alternative first-line drugs could be considered. There were only four reports coded as class I studies, done in partial epilepsies (adults 2, children 1, elderly 1)[Rowan et al. 2005; Chadwick, 1999; Guerreiro et al. 1997; Mattson et al. 1985]. The majority of studies carried out in partial epilepsies and all the studies done in generalized epilepsies were, however, class III (Table 3). Based on the results of the two class I studies and of a class II study done in adult patients with partial epilepsies, a type A recommendation was made for the use of carbamazepine and phenytoin and a type B recommendation was made for the use of valproate. Class I studies found oxcarbazepine effective for the treatment of partial-onset epilepsies in children and adolescents, and gabapentin and lamotrigine effective for the treatment of partial-onset epilepsies in the elderly (type A recommendation)(Table 3). Owing to the lack of high-quality studies, only type C recommendations could be made for the use of first- and second-generation drugs in generalized-onset epilepsies and in idiopathic partial (Rolandic) epilepsies (Table 3). However, at the time of Glauser's systematic review, published evidence for other second-generation drugs (e.g. levetiracetam) was not available. Two pragmatic randomized open trials were also performed to assess the comparative effectiveness of carbamazepine, gabapentin, lamotrigine, oxcarbazepine and topiramate for the treatment of newly diagnosed partial-onset epilepsies (SANAD A)[Marson et al. 2007a] and to assess the effectiveness of valproate, lamotrigine and topiramate for the treatment of newly diagnosed generalized-onset or unclassifiable epilepsies (SANAD B)[Marson et al. 2007b]. In partial-onset epilepsies, lamotrigine was better than carbamazepine, the standard drug treatment, and was found to be a cost-effective alternative. In generalized-onset and unclassifiable epilepsies, valproate was better tolerated than topiramate and more efficacious than lamotrigine, and was confirmed as the drug of first choice. However, in the SANAD studies, drug ‘effectiveness' was based on time to treatment failure, which reflected a combination of drug tolerability and efficacy. In this regard, lamotrigine was better than carbamazepine primarily because it was better tolerated rather than because it more effectively reduced seizures. In addition, because of adverse effects during pregnancy, valproate use in women of childbearing potential should be assessed with caution. In light of these results, drugs to be recommended for partial-onset epilepsies are carbamazepine or lamotrigine in adults, oxcarbazepine in children, and lamotrigine or gabapentin in the elderly. A drug suggested for use in generalized-onset epilepsies is valproate, except for women of childbearing age; in these cases and in other (selected) syndromes the drug should be tailored to the individual patient.
Each of the antiepileptic drugs is mostly used in standard daily doses, which were defined by the results of therapeutic trials done in patients with drug-resistant epilepsy. In contrast, there are several pieces of evidence supporting the use of low daily doses in the majority of patients with epilepsy. These come from more recent monotherapy trials and from studies on the risk of relapse when treatment is discontinued after prolonged seizure remission. In these cases, patients who were satisfactorily controlled with subtherapeutic doses relapsed when treatment discontinuation was completed. In addition, in clinical practice many patients achieve seizure control with low maintenance doses of a given drug. However, the evidence in support of the comparative efficacy of low daily doses of AEDs is scant. For this reason, in the 1990s a pragmatic randomized open trial was started to assess the efficacy of standard versus low (half of standard) doses of anticonvulsant drugs in patients with newly diagnosed epilepsy [Beghi et al. 2005]. In this study, four drugs (those which at the time were commonly used as first treatment) were assessed: carbamazepine, valproate, phenobarbital and phenytoin. A total of 80 children and adults were recruited. The selection of the drug was left to the caring physician's choice. At admission, each patient was randomized to receive standard or low doses (i.e. half the standard dose) of the assigned drug. Although this study was underpowered to detect equivalence because recruitment could not be completed as per protocol, the interim findings are interesting because at the time of closure there were no significant differences between the two treatment strategies in terms of seizure relapse and long-term seizure remission. Treatment modification was, however, more frequent in patients randomized to standard daily doses for the occurrence of adverse events. These findings further confirm the need to tailor the maintenance dose based on the efficacy and tolerability of the assigned drug.
Another important step in the treatment of epilepsy is the strategy to adopt when the first drug fails. The caring physician is confronted with a choice between an alternative monotherapy or an adjunctive treatment (a second compound to be added to the initial drug). The comparative value of the two treatment strategies has been addressed in two randomized trials. In the first [Deckers et al. 2001], patients with previously untreated generalized tonic–clonic and/or partial seizures were randomized to receive carbamazepine or a combination of carbamazepine and valproate at reduced doses. After 1-year follow-up, nonsignificant differences were found between the two treatments in seizure frequency or adverse drug reactions although there was a trend for rates of withdrawal due to adverse effects to be lower in the polytherapy group. In a pragmatic randomized open trial, children and adults in whom a first monotherapy failed were randomized to alternative monotherapy or to add-on therapy. There were no significant differences between the two therapeutic strategies in terms of seizure relapse and discontinuation of the assigned treatment [Beghi et al. 2003]. However, as in the previous trial, alternative monotherapy carried more adverse events than add-on therapy. This finding could be explained in part by the higher daily doses of the drugs used as alternative monotherapy. The safety profile of a drug is thus an important element in the decision to opt for an alternative monotherapy or for an add-on therapy in a patient in whom the first monotherapy failed.
The duration of treatment accompanied by prolonged seizure control is a key issue in the management of epilepsy. A long-term population-based study has shown that 5-year terminal remission (i.e. off-drugs) of epilepsy is 61% [Annegers et al. 1979]. These findings, confirmed by several other subsequent reports, prompt treatment discontinuation when a prolonged period of seizure freedom has occurred. However, only two randomized trials have assessed the effects of AED withdrawal on seizure relapse and patients' performance. In the first trial [MRC Group, 1991], patients randomized to continued treatment showed a 22% relapse at 2years, while patients randomized to slow drug withdrawal had 41% relapse. This differential risk of relapse was maximal between 1 and 2 years and declined thereafter. After 2 years, the risk of subsequent relapse was the same for both treatment groups. The risk of recurrence was also similar in patients who relapsed after withdrawal of AEDs and in those who relapsed while remaining on treatment [Chadwick et al. 1996]. In the second trial [Lossius et al. 2008], 15% of patients randomized to treatment withdrawal and 7% of those randomized to remain on treatment had a relapse at 12 months, a nonsignificant difference. However, compared with the latter, the former improved significantly in their neuropsychological performance. Discontinuation of drug treatment is thus a valuable option in patients with epilepsy who are seizure-free for 2years or longer. However, the decision to withdraw or continue treatment is subject to the calculation of the risk of relapse after treatment stop. In a systematic review of the literature, the relapse rate at two years was found to range from 43% to 65% in adults and from 9% to 39% in children [Specchio and Beghi, 2004]. The factors associated with an increased risk of relapse include the onset of seizures in adolescence, the presence of partial seizures, a documented etiology, and an abnormal EEG at the time of treatment discontinuation. Compared with patients with onset of seizures in childhood, those with onset in adolescence have a 1.8 chance of relapse. The risk of recurrence is 1.5 in patients with a documented etiology of seizures and 1.4 in those with abnormal EEG [Berg and Shinnar, 1994]. This modest increase in risk is still compatible with treatment discontinuation even in patients with adolescent-onset of seizures, symptomatic seizures, and abnormal EEG. In contrast, the various combinations of these risk factors (age at onset, seizure type and etiology, and EEG findings) may help identify selected syndromic patterns, which may be a more precise reference to the decision to stop or to continue the assigned treatment.
In a Cochrane systematic review, Sirven and colleagues  assessed the effects of early discontinuation (i.e. after less than 2-year seizure freedom) versus late discontinuation (i.e. 2-year seizure freedom or longer) on seizure relapse. Early discontinuation was associated with higher recurrence risk than late discontinuation in children, particularly those with partial seizures and/or abnormal EEG. However, the authors concluded that the optimal time of withdrawal was still not clear, there was insufficient evidence to establish when to stop AEDs in children with generalized seizures, early drug withdrawal should be based on the child's underlying syndrome, and there was no evidence to guide timing of withdrawal in adults. Thus, the epilepsy syndrome along with the social, psychological and emotional profile of the patient are all elements contributing to the individualization of the decision to stop treatment in patients with prolonged seizure remission.
The various steps characterizing the treatment of epilepsy (onset, drug choice and daily dose, and duration) are the result of a complex process in which a decision must be taken in light of the outcome of the disease, the factors predicting that outcome, and the efficacy and safety of the available drugs. All these elements help us to understand why the treatment of epilepsy across its different stages cannot be standardized; it must be largely individualized and subjected to a comprehensive evaluation of the individual case.
The author has nothing to declare with reference to the contents of this manuscript.