In the past, almost all children with a seizure of any type, febrile or afebrile, were placed on long-term therapy with antiepileptic drugs (AEDs). This was based on several assumptions: first, that almost all children with an isolated seizure would go on to have more seizures (1,2); second, that seizures, even brief ones, could cause brain damage and lead to progressively intractable epilepsy (1–3); third, that AEDs were not only effective but also safe and that treatment was associated with only minimal morbidity; finally, that seizures “beget” seizures, and that early AED therapy not only prevented seizures but also somehow altered the natural history of the disorder and prevented the development of “chronic” epilepsy (3–6). We now know that these assumptions are not true. Research has provided information that has altered the way physicians think about seizures, their consequences, and the drugs used to treat them. The decision whether and for how long to treat a child with AEDs must be weighed against the possible risks of that treatment and must take into account the large body of data that has accumulated: namely, that many children with a single seizure do not go on to experience further seizures (7–20); that many children with epilepsy ultimately go on to become seizure free (21–24); that most seizures are brief, and that even prolonged seizures rarely cause brain damage unless they are associated with an acute neurologic insult (25–28); and that antiepileptic medications can cause untoward effects (18,29–31).

The decision of whether to initiate treatment in a child with one or more seizures must balance the risks and benefits of treatment in each case (18). Similarly, the patient who is seizure free on medications for some time must weigh the risks of possible seizure recurrence if medications are withdrawn against the risks of continuing long-term AED therapy. This chapter reviews the data relevant to these decisions. The data on the probability of seizure recurrence following a first unprovoked seizure are presented. Next, the issue of withdrawing antiepileptic drugs in children with epilepsy who are seizure free for 2 or more years is considered. This is followed by a review of the risks of not treating (the risks of subsequent seizures) and the morbidity of therapy. Finally, recommendations for a therapeutic approach to children with seizures are outlined.

RECURRENCE RISK FOLLOWING A FIRST UNPROVOKED SEIZURE

An understanding of the natural history of children who present with a first unprovoked seizure is necessary in order to develop a rational approach to their management. Over the past two decades, there have been many studies that have attempted to address this issue (7–20). For purposes of this discussion, a first unprovoked seizure is defined as a seizure, or flurry of seizures all occurring within 24 hours, in a patient over 1 month of age with no prior history of unprovoked seizures (18,32).

The reported overall recurrence risk following a first unprovoked seizure in children varies from 27% to 71% (7–20). Studies that identified the children at the time of first seizure and carefully excluded those with prior seizures report recurrence risks of 27% to 44% (7–11,19). Studies that recruited subjects later, either retrospectively or from electroencephalogram (EEG) laboratories, but excluded those with prior seizures, report slightly higher recurrence risks of 48% to 52% (14,15). Lastly, studies that included children who already had recurrent seizures at the time of identification report the highest recurrence risks, 61% to 71% (12,13). Once methodological issues and differences in the distribution of risk factors among different studies are taken into account, the results are fairly consistent (9). The majority of recurrences occur early, with approximately 50% of recurrences occurring within 6 months and more than 80% within 2 years of the initial seizure (7–20). Late recurrences are unusual.

Similar predictors of recurrent seizures were found in the majority of the studies despite variations in methodology and in subject selection (7–20). Factors that are associated with a differential risk of recurrence include the etiology of the seizure, the electroencephalogram (EEG), whether the first seizure occurred in wakefulness or sleep, and seizure type. Factors not associated with a change in the recurrence risk include age of onset and the duration of the initial seizure. The report of a family history of seizures in a first-degree relative is of unclear significance, with conflicting results in the various studies. Risk factors for seizure recurrence from our large prospective study (8,20) are shown in Table 38.1.

Selected individual risk factors are discussed in the following subsections.

Etiology

The International League Against Epilepsy (ILAE) guidelines for epidemiologic research (32) classify seizures as acute symptomatic, remote symptomatic, cryptogenic, or idiopathic. Acute symptomatic seizures are those associated with an acute insult such as head trauma or meningitis. Remote symptomatic seizures are those without an immediate cause but with an identifiable prior brain injury such as major head trauma (loss of consciousness greater than 30 minutes, depressed skull fracture, or intracranial hemorrhage), meningitis, encephalitis, stroke, or the presence of a static encephalopathy, such as mental retardation or cerebral palsy, which are known to be associated with an increased risk of seizures. Cryptogenic seizures are those occurring in otherwise normal individuals with no clear etiology. Note that factors such as sleep deprivation are considered trigger factors but do not change the classification of the seizure, as they would be associated with seizures only in susceptible individuals. Until recently cryptogenic seizures were also called idiopathic. In the new classification, the term idiopathic is reserved for seizures occurring in the context of the presumed genetic epilepsies, such as benign Rolandic and childhood absence epilepsy (32–34), but many papers still refer to cryptogenic seizures as idiopathic.

Children with a remote symptomatic first seizure have higher risk of recurrence. In one large prospective study with mean follow-up over 10 years, 44 (68%) of 65 children with a remote symptomatic first seizure had a recurrence, compared with 127 (37%) of 347 children with a cryptogenic/idiopathic first seizure (P < .001) (8,19). Comparable findings are reported in other studies (9–11,14,15).

Electroencephalogram

The EEG is an important predictor of recurrence, particularly in cryptogenic cases (8–11,14,17). Epileptiform abnormalities are more important than nonepileptiform ones, but any EEG abnormality increases the recurrence risk in cryptogenic cases (8). In our study, the risk of seizure recurrence by 5 years for children with a cryptogenic first seizure was 27% for those with a normal EEG, 44% for those with nonepileptiform abnormalities, and 62% for those with epileptiform abnormalities (7,8). In our data (8,35), any clearly abnormal EEG patterns, including generalized spike-and-wave, focal spikes, and focal or generalized slowing, increased the risk of recurrence, while Camfield et al (14) report that only epileptiform abnormalities substantially increase the risk of recurrence. Hauser et al (10) state that only generalized spike-and-wave patterns are predictive of recurrence, but they studied mostly adolescents and adults and thus would have not included many children with centrotemporal spikes (benign rolandic epilepsy), which is the most common focal spike pattern found in studies focusing on children with a first seizure (8,31,14). In children with a cryptogenic first seizure, the EEG appears to be the most important predictor of recurrence (8,9). Based on these data, a recent practice parameter of the American Academy of Neurology recommended that an EEG be considered a standard part of the diagnostic evaluation of the child with a first unprovoked seizure (17).

TABLE 38.1

Sleep State at Time of First Seizure

Whether the initial seizure occurs while the child is awake or asleep is associated with a differential recurrence risk, particularly in cryptogenic cases (8). In our series, the 5-year recurrence risk was 53% for children whose initial seizure occurred during sleep, compared with 36% for those whose initial seizure occurred while awake (P < .001) (8). On multivariable analysis, etiology, the EEG, and sleep state were the only significant predictors of outcome. The group of children with a cryptogenic first seizure occurring while awake and a normal EEG had a 5-year recurrence risk of only 21%. If seizures did recur, they usually recurred in the same sleep state as the initial seizure.

Seizure Classification

Some, though not all, studies indicate that the risk of recurrence following a partial seizure is higher than that following a generalized seizure (9). However, partial seizures are more common in children with a remote symptomatic first seizure and in children with a cryptogenic first seizure who have an abnormal EEG (7). Once the effects of etiology and the EEG are controlled for, partial seizures are not associated with an increased risk of recurrence (7,8,10). In the meta-analysis, no clear association between seizure type and recurrence risk could be found (9).

Family History

At present, there are insufficient data to determine whether a positive family history of epilepsy is a risk factor for recurrence. Although one study, primarily focused on adults, found a substantially increased risk of recurrence in those with a positive family history of epilepsy (10), others have failed to find a major effect (8,14). In our study, family history was important only in children with a cryptogenic first seizure who also had an abnormal EEG. This type of patient constitutes a very small fraction of the population (7). These mixed results suggest that the additional risk of a positive family history, if present, will turn out to be small or else limited to specific subgroups.

Duration of First Seizure or Status Epilepticus

The duration of the first seizure does not affect the risk of recurrence (8–10). This is true whether one analyzes it as a continuous variable or separates the children into those who had status epilepticus and those who had a briefer seizure (8). Most cases of status occur as the initial seizure. In our series of 407 children, 48 (12%) had status epilepticus as their seizure, but only 7 (4%) of the 171 children with recurrent seizures recurred with status. Although the occurrence of status epilepticus as the first seizure did not alter recurrence risks, a recurrence was more likely to be prolonged. Five (21%) of 24 children with recurrent seizures whose initial seizure was status had an episode of status as their second seizure, compared with 2 (1%) of the 147 children with recurrent seizures whose first seizure was brief (P < .001). None of these experienced any sequelae (8). Remission rates were not different in those who presented with an episode of status epilepticus (36).

Age at First Seizure

The majority of studies, in both children and adults, have not found age at first seizure to alter the risk of recurrence (7–11,14–16). This was true whether age was analyzed as a continuous variable or broken up into several age ranges. The only exception to this is the National Collaborative Perinatal Project, which found an increased risk of recurrence in children under age 2 with focal motor seizures (13). At present the preponderance of available data shows that the age at time of first seizure does not affect the risk of recurrence following a first unprovoked seizure.

Treatment Following a First Seizure

In observational studies such as those just discussed, whether children were treated or not after their first seizure did not alter the recurrence risk (8,10). However, these were not randomized treatment trials. The physicians presumably treated those children whom they thought had a high risk of recurrence. Following an initial seizure, patients are often started on a small dose of medication, and compliance may be lax. Randomized clinical trials comparing AED therapy with placebo following a first unprovoked seizure in children and adults have found that AED therapy can reduce the risk of a second seizure by half (16,37). However, with longer follow-up, there was no difference between the two groups in terms of the probability of achieving remission (6,38). Based on a review of the evidence, the American Academy of Neurology practice parameter on the treatment of the child with a first unprovoked seizure concluded that treatment does not prevent the development of chronic epilepsy or alter long-term outcome and should be reserved for those cases where the risks of a second seizure outweigh the morbidity of AED therapy (18). Similar conclusions—namely, that treatment reduces recurrence risk but does not alter long-term prognosis—were reached in the recently published American Academy Practice Parameter for treatment following a first seizure in adults (39).

CONCLUSIONS

Knowing these predictors and the recurrence risks, the child with a first unprovoked seizure presents an interesting dilemma. The likelihood that it will be an isolated event that will not repeat itself must be weighed against whether it is the first of many attacks. A thorough evaluation of the patient, including a detailed history and appropriate laboratory studies such as an electroencephalogram (EEG), is indicated regardless of whether AED therapy is started (17). Factors such as the seizure type, family history of seizures, and the possible etiology of the seizure must be ascertained. Of particular importance is a careful history of prior events that may be seizures. Many children who first come to medical attention as a result of a convulsive episode are found to have had prior nonconvulsive episodes of absence or complex partial seizures that were not recognized as such by the family (8,17). These children clearly fall into the category of newly diagnosed epilepsy and not first seizure.

The majority of children with a first unprovoked seizure do not have additional unprovoked seizures. Children with a cryptogenic first seizure and a normal EEG have a particularly favorable prognosis. There are small subgroups of children with multiple risk factors who do have a high risk of recurrence. However, although AED therapy will reduce the likelihood of further seizures, there is no evidence that it alters the long-term prognosis (18,38). In particular, the data from randomized clinical trials and large epidemiologic studies indicate that delaying therapy will neither alter the response rate to AEDs nor adversely affect the probability of attaining remission (3,5,6,18,23,38,39). Finally, recent data suggest that children who present with a first unprovoked seizure are very unlikely to develop medically refractory epilepsy even if seizures do recur (19,20). The decision to treat or not after a first seizure must be based on the relative risks and benefits of therapy compared with the risks of further seizures. This risk benefit assessment is discussed at the end of this chapter. The authors rarely treat children with a first unprovoked seizure.

WITHDRAWING AEDs IN CHILDREN WITH EPILEPSY WHO HAVE BEEN SEIZURE FREE FOR 2 OR MORE YEARS

The available data indicate that children who are seizure free on medication for 2 or more years have a very high likelihood of remaining in remission on medication (40). In selected populations, withdrawal may be feasible after an even shorter seizure-free interval (40–43). How long should a child be maintained on medication before the attempt is made to withdraw it? This decision will be influenced by a variety of factors, including the probability of remaining seizure free after withdrawal in a given patient, the potential risk of injury from a seizure recurrence, and the potential adverse effects of continued AED therapy.

The majority of children who are seizure free on medications for at least 2 years will remain seizure free when medications are withdrawn. A large number of well-designed studies involving more than 700 children have been done over the past 20 years (40,44–55). The overall results have been very similar. Between 60% and 75% of children with epilepsy who have been seizure free for more than 2 (40,44,45,50–54) or 4 (45–49) years on medications remain seizure free when antiepileptic medication is withdrawn. Furthermore, the majority of recurrences occur shortly after medication withdrawal, with almost half the relapses occurring within 6 months of medication withdrawal and 60% to 80% within 1 year (40,54,55). These studies are supported by a follow-up of patients for 15 to 23 years after withdrawal of medication (48,56). Although late recurrences do occur, they are rare (48,54,56). In a recent randomized study, the increased risk of relapse following AED withdrawal occurred only in the first 2 years after AED withdrawal. The rate of late recurrences was the same in those who remained on AED therapy and those whose AEDs were discontinued (49).

The important question is: Can one identify risk factors such as etiology, age of onset, type of seizure, EEG features, or the specific epilepsy syndrome, that will enable one to identify subgroups of children with an even better prognosis and subgroups with a much less favorable prognosis for maintaining seizure remission off medication? There is much less consensus in this area. A discussion of the potential risk factors that have been looked at and their possible significance is presented next.

Etiology and Neurologic Status

In general, children with epilepsy associated with a prior neurologic insult have a lower chance of becoming seizure free in the first place than do children with cryptogenic epilepsy (21,22,24). In children with remote symptomatic epilepsy who are seizure free on medication, most studies indicate a higher risk of recurrence following discontinuation of medication than in children with cryptogenic epilepsy (45–48,50,51,54). In a recent meta-analysis of this literature, the relative risk of relapse in those with remote symptomatic seizures was 1.55 (95% CI 1.21–1.98) (55). However, almost half of these children will remain seizure free after withdrawal of medication (46,54,55). Furthermore, even within this group, one can identify subgroups with favorable and unfavorable risk factors (54).

Age of Onset and Age at Withdrawal

Age of onset above 12 years is associated with a higher risk of relapse following discontinuation of medications (40,43,46,50–52,54,55). In our data, this was the single most important risk factor for recurrence (relative risk 4.24, 95% CI 2.54–7.08). A meta-analysis (55) also found adolescent-onset seizures to be associated with a higher risk of recurrence than childhood-onset (relative risk 1.79, 95% CI 1.46–2.19). There is some controversy as to whether a very young age of onset (under 2 years) may be a poor prognostic factor (46). In our data, a young age of onset was associated with a less favorable prognosis only in those with remote symptomatic seizures, and was associated with more severe neurologic abnormalities (54). As most childhood epilepsy is readily controlled with AED therapy, the age at withdrawal of AEDs will be highly correlated with the age of onset. However, the age at AED withdrawal does not appear to be important once age of onset is taken into account. In particular, there is no evidence that discontinuation of AEDs during the pubertal period is associated with a higher risk of recurrence (40,44,45,54).

Duration of Epilepsy and Number of Seizures

These two variables are closely interrelated. A long duration of epilepsy may increase the risk of recurrence, though the magnitude of the effect is small (40,43,47,48). One study also reported that having more than 30 generalized tonic–clonic seizures was associated with a high risk of recurrence after discontinuation of therapy (46). In a community-based practice, most children will be easily controlled within a short time after therapy is initiated, so these factors will rarely be important.

Seizure Type

Studies investigating the effect of seizure type on the risk of recurrence after medication withdrawal in children have produced inconsistent results. Children with multiple seizure types have a poorer prognosis (47). The data regarding partial seizures are conflicting (40,44–55). At this time, it is not clear that any specific seizure type is associated with an increased risk of recurrence following discontinuation of medication.

EEG

In several studies (40,41,44–46,54,55), the EEG prior to discontinuation of medication was one of the most important predictors of relapse in children with cryptogenic epilepsy. However, the specific EEG abnormalities of significance varied across studies. Two other studies found no correlation between the EEG and outcome (47,51). A meta-analysis found that an abnormal EEG prior to AED withdrawal was associated with a relative risk of relapse of 1.45 (95% CI 1.18–1.79) (55). The preponderance of evidence indicates that an abnormal EEG is associated with an increased recurrence risk in children with cryptogenic or idiopathic epilepsy.

The EEG obtained at the time of initial diagnostic evaluation may also have predictive value. Certain characteristic EEG patterns associated with specific epileptic syndromes, such as benign rolandic epilepsy or juvenile myoclonic epilepsy, provide additional prognostic information (24,33,34,54). Changes in the EEG over time may also have prognostic value (41,44).

Epilepsy Syndrome

Epilepsy syndromes are known to be associated with a differential prognosis for remission (24,33,34). Regrettably, there is little information as to the effect of the specific epilepsy syndrome on the risk of relapse following AED withdrawal. The majority of studies of AED withdrawal have not provided information by epilepsy syndrome. The results from our data (54) are shown in Table 38.2. Overall, patients with both idiopathic and cryptogenic epilepsy syndromes have similar prognoses. However, specific syndromes are associated with a differential risk of relapse. Patients with benign rolandic epilepsy have a particularly favorable prognosis, even if their EEGs are still abnormal, whereas all four of our patients with juvenile myoclonic epilepsy relapsed. Clearly, future studies will have to focus more on the role of the specific epilepsy syndrome in guiding therapy, both in terms of initiating and discontinuing therapy and in the selection of appropriate treatment.

TABLE 38.2

Type of Medication

The majority of studies have not found that the specific AED used affects recurrence rates. One well-designed randomized study in adults (53) suggests that the risk of recurrence may be higher in those treated with valproate than in those treated with other medications. The significance of this finding remains unclear. It may be related to the ability of valproate to normalize generalized spike–wave abnormalities, thus making the subject appear to be at lower risk than is actually the case. At present, there is insufficient evidence to justify basing the decision to continue or withdraw AEDs on the type of AED the patient is taking.

The serum drug level does not seem to have a great impact on recurrence risk. Children who have not had seizures for several years often have “subtherapeutic” levels, and few have toxic levels. Available studies show little or no correlation between drug level prior to discontinuation and seizure recurrence and outcome (44), or a very modest effect (46).

Duration of Seizure-Free Interval

The chances of remaining seizure free after medication withdrawal is similar whether a 2-year (40,44,45,50–54) or 4-year (40,45–49) seizure-free interval is used. One study that evaluated seizure-free intervals of 1 or more years did find that a longer seizure-free period was associated with a slightly lower recurrence risk (45). Note that among children who are 2 years seizure free but remain on medication, approximately 3% to 5% will experience another seizure in the third or fourth year of treatment (21). More recent studies that have utilized a seizure-free interval of 1 year or less have reported higher recurrence risks (40–43). The risk of relapse after a 1-year remission, compared with a longer seizure-free interval, is also higher in patients who continue on AEDs (21).

Remission Following Relapse

The majority of patients who relapse after medication withdrawal will reattain remission after AEDs are restarted, though not necessarily immediately (40,56–58). The prognosis for long-term remission appears to be primarily a function of the underlying epilepsy syndrome. A recent randomized study of medication withdrawal found that the prognosis for seizure control after recurrence in patients with previously well-controlled seizures was no different in those who were withdrawn from AED therapy and relapsed and those who relapsed while remaining on AED therapy (57).

Withdrawal of AEDs Following Successful Epilepsy Surgery

Epilepsy surgery is an increasingly used treatment modality in children with localization-related, medically refractory epilepsy (59–61). In the past, these children tended to stay on medication. Increasingly for those who are seizure free, efforts are being made to taper medications in a subset where it is believed that the epileptogenic zone has been fully resected. Successful withdrawal of AEDs has been reported in children following successful respective surgery (62,63). Generally, the risk of AED withdrawal in this group of children is similar to those with remote symptomatic epilepsy who are seizure free on their medication (40,54,55). It is therefore reasonable to consider withdrawing medications in children who have had successful epilepsy surgery. The duration of the seizure-free interval needed following surgery is unclear. In principle, if one has eliminated the focus, the need to wait is unclear. However, sometimes the surgery disrupts the system with transient improvement but not full long-term remission. The practice has, therefore, been to wait at least 2 years before considering full withdrawal of medication (62,63). This discussion focuses on full withdrawal of all AEDs, as reducing the number of AEDs in these children is already standard.

RISKS OF NOT TREATING OR OF DISCONTINUING AEDs

Stay updated, free articles. Join our Telegram channel

Join