General Aspects of Prognosis

The prognosis for children with epileptic seizures depends largely on the type of epilepsy syndrome responsible for the attacks. Throughout this book, the concept that different epilepsy syndromes have different outcomes has been emphasized, so a global study of prognosis for childhood epilepsies mainly reflects the distribution of the various epilepsy syndromes in the studied samples. However, an overview of the prognostic aspects of the epilepsies of children in general may be of interest for several reasons. First, not all of the epilepsies can be separated into definite syndromes, so a significant proportion of cases remain unclassified. In such cases, some general method for assessing prognosis is required. Second, global studies of prognosis provide at least some idea of the magnitude of the problems raised by the convulsive disorders of children and of the needs of the community with regard to these individuals. Finally, studies of the global prognosis of childhood epilepsy have shown that certain factors (e.g., frequency of seizures, presence or absence of neurologic or mental deficits, type of seizure, and duration of seizures) have a strong correlation with the outcome of epilepsy. Such factors have a predictive value, not only for unclassified cases of epilepsy but also for some cases belonging to well-defined syndromes. The ready applicability of the factors of prognosis justifies their study.

For many years, the prognosis of epilepsy was considered poor. Rodin (1989) characterized epilepsy as a chronic condition with an average rate of long-term control of 32% in children and 28% in adults. This gloomy outlook was probably due to the patients who were studied, as they were from institutions rather than the general population, as well as to multiple other biases (Sander and Sillanpää, 1998). Indeed, the discrepancy between the lifetime prevalence rate for epilepsy of 2% to 5% in most epidemiologic surveys (Hauser et al., 1993) and the concomitant prevalence rate for active epilepsy of about 0.5% implies that most patients have a favorable outcome.

As Robinson (1984) pointed out, use of evidence from the literature is particularly difficult in studies of prognosis. The differences among studies are often large, and they reflect the manner in which selection biases influence the patient population of particular series and the different proportion of various syndromes in different series. In addition, few studies are concerned specifically with childhood epilepsy, and several series have included mixed populations of adults and children, so some caution must be exercised in applying the results to children. Furthermore, the diagnosis of epilepsy can be difficult, and the inclusion of nonepileptic seizures and a lack of recognition of “minor” attacks are both common. Both can influence the results, as can different rates of reporting by patients or families (Sander et al., 1998). Several investigators have emphasized the importance of the temporal aspects of seizure disorders on relapse and remission rates (Annegers et al., 1986; Elwes et al., 1984, 1985; Elwes and Reynolds, 1981). Patients with long-standing epilepsy are likely to continue to have seizures, whereas “fresh” patients with only two or few seizures have seizures that are much more likely to remit. The results, therefore, vary with the proportion of this type of patient. Moreover, chronic cases are referred more often to specialized units, whereas “fresh” cases are often treated locally. As a result, series from specialized centers give a more pessimistic view of the prognosis.

Another reason for the differences in the results is the variable interval separating the first seizure from entry into the studies. When this interval is prolonged, a proportion of the original patients may have experienced another seizure, and these individuals are therefore not suitable for inclusion. This leads to the underestimation of the true recurrence rate of a first seizure. In one adult series (Hirsch et al., 1991a), the rate was 15% for patients who entered the study within 8 weeks of their first seizure, compared with 50% in those who registered within 4 weeks.

PROGNOSIS FOR SEIZURE RECURRENCE AND REMISSION

The prognosis of seizures is of primary importance because the occurrence of attacks influences all aspects of the individual’s quality of life. The problem presents differently in patients who have experienced a first epileptic seizure and those with established recurrent
epilepsy. These two situations are considered successively.

Recurrence Rate after a First Seizure

The evidence in this regard is conflicting, with an estimated recurrence risk that varies from 27% to 81% (Hopkins et al., 1988) (see Chapter 15).

In series including adults or children, the rate is generally lower in series from the general population than it is in hospital-based series (Hopkins et al., 1988). The issue is complex because time factors (i.e., how early after the event the patient is seen), the circumstances of occurrence (i.e., how certain is the absence of precipitating factors), antecedents (personal and familial), and the type of seizure and syndrome are involved. This section considers only those cases in which the first seizure is of the generalized or localized motor type, is unprovoked, and does not take place at the time of an acute disorder.

Recent studies focusing exclusively on children have shown that the risks vary with the type of seizure and the presence or absence of neurologic and electroencephalographic (EEG) abnormalities (Berg and Shinnar, 1991; Shinnar et al., 1990; Boulloche et al., 1989; Arts et al., 1988; Camfield et al., 1985b; Hirtz et al., 1984). Recurrence was seen in 87 (52%) of 168 children after an average of 32 months in one series (Camfield et al., 1985b). In this series, the risk was highest for patients with sylvian (rolandic) seizures (70% when the EEG was normal and 97% when it was abnormal); next highest for patients with other types of partial seizures, whether simple or complex (50% to 83% depending on type, EEG, and results of neurologic examination); and lowest for patients with generalized tonic-clonic seizures (30% when no EEG or neurologic anomalies were found and 63% when both neurologic and EEG abnormalities were present). Roughly similar figures have been found in other studies (Berg and Shinnar, 1991; Shinnar et al., 1990; Boulloche et al., 1989; Arts et al., 1988), but lower figures (e.g., 16% at 1 year and 27% at 3 years [Annegers et al., 1979a]) and higher figures (e.g., 67% and 78% at 1 and 3 years, respectively [Hart et al., 1990]) have been recorded in mixed adult and children series. Recent work has shown that EEG results can reliably identify children with very high (those with paroxysmal abnormalities) or very low recurrence rates (those with normal EEG) (Hirtz et al., 2000; Shinnar et al., 1994).

Discrepancies among the figures that are given likely result from sampling and methodologic differences; some patients with an isolated seizure may never be seen in a specialized center or even by a doctor (Costeff and Avni, 1982), which may result in an apparent increase in recurrence rate. On the other hand, some series may include children with nonepileptic seizures, thus lowering the recurrence rates because these are not likely to be followed by continuing epilepsy.

Notwithstanding such methodologic difficulties, a sizable proportion of children with a first seizure (approximately 40% to 60%) do not experience a recurrence over the duration of follow-up. Moreover, the occurrence of a second episode, even though this arbitrarily defines epilepsy, does not signify that the child will develop a chronic incapacitating disorder.

The relatively high incidence of isolated epileptic seizures in adolescents, especially those of a partial motor or complex type, is discussed in Chapter 15. When a single seizure or a cluster of seizures occurs in an adolescent, the overall prognosis is favorable, with a likelihood for remaining seizure free of 67% after 5 years and 76% after 10 years. With simple partial seizures and a normal EEG, 100% of patients remain seizure free after 3 years. The figure is slightly lower (89%) for patients with complex partial seizures and a normal EEG (King et al., 1999; Loiseau and Orgogozo, 1978).

Approximately 20% of children with rolandic seizures have only one attack, as retrospective studies have shown (Ambrosetto et al., 1987; Beaussart and Faou, 1978; Loiseau and Beaussart, 1973; Beaussart, 1972).

The aforementioned figures have an obvious impact on the decision to treat. The decision should be discussed with the patients and parents because no firm evidence-based decision is possible when the clinician is facing the individual patient. Certainly, abstaining from therapy is reasonable when the risk of recurrence is low.

Outcome for Seizures in Patients with Confirmed Epilepsy

For the purpose of epidemiologic study, epilepsy is defined as the occurrence of at least two spontaneous (unprovoked) seizures that do not supervene in the same episode and that are not related to fever or an acute disorder. This definition is purely operational. However, patients who experience two seizures may have a higher risk of recurrence than do those with a first seizure, but they do not necessarily develop chronic epilepsy. Of those patients who experienced a second seizure, 79% went on to have at least a third seizure, even though they usually received antiepileptic
agents (Camfield and Camfield, 2000). In a cohort of adults, Hauser and Hersdoffer (1990) and Hauser et al. (1990) noted that the recurrence rate after a second seizure was 65%, and 63% of those with a third seizure also experienced a fourth. In fact, including patients who had already experienced a second seizure in their study of the recurrence risk in patients who had experienced a first seizure resulted in only a minor increase in relapse rate (Hauser et al., 1990).

Remission Rate

Most studies of seizure outcome deal with treated epilepsy, even though the highly variable modalities of therapy and patient compliance are almost universally unreported. Figures for the recurrence rate in newly diagnosed epilepsy vary between 35% and 57% in the literature (Hart et al., 1990; Sander and Shorvon, 1987; Annegers et al., 1979).

In series of childhood epilepsy, more than half the patients were seizure free after a 10-year or longer follow-up. Actual figures at that point range between 51% (Sillanpää, 2000) and 82% (Oka et al., 1989). In a long (12 years), prospective, follow-up study of 119 children, Brorson and Wranne (1987) found an overall terminal remission rate of 64%. The 3-year terminal remission rate was 89% if the child’s neurologic status was normal but 49% if the patient had “mental retardation” and/or abnormal neurology. In a study including both children and adults (Annegers et al., 1979), the remission rates were 42%, 61%, and 70% after follow-ups of 1 year, 10 years, and 15 years, respectively. In several long-term, prospective studies, Sillanpää (1990, 2000) and Sillanpää et al. (1998a) found 5-year terminal remission rates of 48% at 10 years, 56% at 20 years, 60% at 30 years, and 70% at 40 years. Interestingly, these studies included a sample of children with epilepsy that was typical of the child population in a well-defined, representative catchment area of Finland. About half the patients who were in remission were not on medication. The remission rate depended on the etiology of the seizures, with the 5-year terminal remission rates for idiopathic, cryptogenic, and remote symptomatic seizures being 95%, 68%, and 45%, respectively (Sillanpää, 2000). Another recent prospective study (Cockerell et al., 1997) found a 3-year terminal remission rate of 66% and a 5-year rate of 46% in children younger than 16 years after a 9-year follow-up. However, other studies have found a less favorable outcome. Ellenberg et al. (1984) found that 344 (67%) of their 513 children with epilepsy, excluding those with febrile convulsions and seizures associated with acute systemic disorders, still had seizures at the age of 7 years. Methodologic differences, especially in the criteria of inclusion and duration of follow-up, are probably important (Beghi and Tognoni, 1988; Goodridge and Shorvon, 1983a).

Excluding children with obvious associated brain damage, the overall prognosis for the seizures in those with “ordinary” uncomplicated epilepsy (Camfield and Camfield, 2000) is good in 70% to 80% of cases (Sillanpää et al., 1998a; Sander, 1993). Even lesional epilepsies can remit. Huttenlocher and Hapke (1990) followed 155 children, 145 of whom were not operated on, with refractory epilepsy that was not due to a progressive central nervous system disorder or tuberous sclerosis; 61% had mild to moderate retardation, and 39% were of borderline intelligence. Good seizure control in these children increased by roughly 4% yearly starting about 4 years after onset of epilepsy, and only one-fourth of those followed for 18 years or more continued to have more than one seizure per year. Even 1.5% of the children with mental retardation achieved control during each year of follow-up, but 70% still had seizures after 18 years. Similar results have been reported by Goulden et al. (1991). However, achieving control of such severe epilepsies by drug treatment takes often inordinately long periods of time, so surgery should be considered early when the seizures continue despite therapy.

Factors associated with structural brain pathology and the propensity toward intense seizures, such as age at onset and frequency of seizures, indicate a higher likelihood of persistence (Table 22.1). A symptomatic cause of epilepsy is especially ominous. Conversely, a rapid response to therapy is an important predictor of lasting remission with an odds ratio for remission of 2.22 to 1. Indeed, in a prospective, multifactorial study of 792 adults and children with seizures, McDonald et al. (2000) found that the single most important prognostic factor for remission was the number of seizures in the initial period of the disease. However, the number and the resistance of seizures might be related to the etiology. Complex partial seizures or atonic attacks herald a difficult epilepsy with odd ratios of remission of 0.26 to 1 and 0.28 to 1, respectively (Sillanpää, 2000).

The significance of the EEG as a predictor of outcome has been variably interpreted (Shinnar et al., 1994; Theodore et al., 1984b; Todt, 1984; Emerson et al., 1981; Rowan et al., 1980), which is hardly surprising because the tracings are not obtained under standard conditions and they represent only brief samples of the total EEG activity. The likelihood of obtaining patterns of prognostic value, if any are present,
varies with the number of recordings and the recording conditions. A normal EEG is a favorable omen (Shinnar et al., 1994). Sleep tracings and special recording techniques, such as long-duration tracings in a laboratory, or free conditions, especially when they are coupled with video recording, improve the value of EEG. However, no absolute relationship between the EEG and the recurrence of seizures exists.

TABLE 22.1. Probability of recurrences of seizures with the presence or absence of certain factors

Factor	Low Probability	High Probability
Abnormalities on neurologic examination^a	Absent	Present
Intellectual function^b	Normal	Disturbed
Demonstrable brain lesion^b	Absent	Present
Number of seizures	One	Multiple
Age at onset^a	≥3-4 yr	<3 yr
Duration of uncontrolled seizures^b	Brief	Long
Response to therapy^b	Rapid	Slow/absent
Frequency of seizures^c	Low	High
Tonic and/or atonic^a seizures, episodes of status, complex partial seizures^c	Absent	Present
Electroencephalogram^a	Normal or normalized rapidly with treatment	Persistently abnormal
Treatment	Regularly taken	Irregular
^aSignificant factor. ^bMajor factor. ^cSignificance uncertain.

An early age at onset has been associated with a poor outcome in most studies (Sillanpää, 2000; Todt, 1984; O’Leary et al., 1981) but not in that of Ellenberg and Nelson (1978), whose study was limited to children younger than 7 years. Epilepsies with onset in infancy and early childhood often have a severe course. Overall, a combination of favorable factors, especially the absence of brain damage, normal intelligence and neurologic examinations, and a rapid response to therapy, predicts an absence of seizures in the long term with a probability of 90% to 95% (Sillanpää, 2000).

Only gold members can continue reading. Log In or Register to continue