Predicting antiepileptic drug response in children with epilepsy

Abstract

In clinical practice, after diagnosis and when treatment has begun, it is important to predict as soon as possible which children will become seizure-free and which are likely to develop medically intractable seizures. This article summarizes factors predicting seizure remission in childhood-onset epilepsy treated with antiepileptic drugs (AEDs). Sustained seizure remission can be expected in over 90% of idiopathic epilepsies of childhood and in neurologically normal children with epilepsy having infrequent seizures showing early remission after starting treatment with AEDs. Even in the presence of symptomatic etiology of epilepsy – focal seizures and syndromes; high seizure frequency prior to or during treatment; seizure clustering; and poor or delayed response to first adequate drug therapy – up to 60% of children with treated epilepsy are able to enter long-term remission. However, remission can be expected in only 30% or less of those with catastrophic epilepsies of childhood.

Keywords::

• drug-resistant epilepsy
• epilepsy
• future epilepsy management
• prediction
• remission
• seizure freedom

Medscape: Continuing Medical Education Online

This activity has been planned and implemented in accordance with the Essential Areas and policies of the Accreditation Council for Continuing Medical Education through the joint sponsorship of Medscape, LLC and Expert Reviews Ltd. Medscape, LLC is accredited by the ACCME to provide continuing medical education for physicians.

Medscape, LLC designates this Journal-based CME activity for a maximum of 1 AMA PRA Category 1 Credit(s)™. Physicians should claim only the credit commensurate with the extent of their participation in the activity.

All other clinicians completing this activity will be issued a certificate of participation. To participate in this journal CME activity: (1) review the learning objectives and author disclosures; (2) study the education content; (3) take the post-test and/or complete the evaluation at http://www.medscape.org/journal/expertneurothera (4) view/print certificate.

Release date: June 8, 2011; Expiration date: June 8, 2012

Learning objectives

Upon completion of this activity, participants should be able to:

• • Describe factors predating epilepsy that predict seizure remission in patients with childhood-onset epilepsy treated with AEDs

• • Describe seizure-related factors that predict seizure remission in patients with childhood-onset epilepsy treated with AEDs

• • Describe treatment-related and other factors that predict seizure remission in patients with childhood-onset epilepsy treated with AEDs

Financial & competing interests disclosure

EDITOR

Elisa Manzotti,Editorial Director, Future Science Group, London, UK

Disclosure:Elisa Manzotti has disclosed no relevant financial relationships.

CME AUTHOR

Laurie Barclay, MD,Freelance writer and reviewer, Medscape, LLC

Disclosure:Laurie Barclay, MD, has disclosed no relevant financial relationships.

AUTHORS

Matti Sillanpää,Departments of Public Health and Child Neurology, University of Turku, Turku, Finland

Disclosure:Matti Sillanpää has disclosed no relevant financial relationships.

Dieter Schmidt,Epilepsy Research Group, Berlin, Germany

Disclosure:Dieter Schmidt has disclosed no relevant financial relationships.

Figure 1. Potential predictive factors involved in seizure outcome of children with epilepsy.

AED: Antiepileptic drug.

The ultimate goal of epilepsy treatment is seizure freedom without any disturbing side effects. Although seizure freedom is seen more often in childhood-onset epilepsy than in epilepsy starting in adults, not all children become seizure-free. In clinical practice, it is of importance to promptly predict, following diagnosis and starting treatment, which children will become seizure-free and which are likely to develop medically intractable seizures. Predicting which child has a good chance of becoming seizure-free is reassuring for the patient and parents or caregivers. Early prediction of poor seizure outcome allows physicians to adequately inform the parents or caregivers and is useful in helping plan allocation of resources for medical support including surgery, if needed. This article focuses on antiepileptic drug (AED) response in terms of seizure outcome reported in recent population-based studies of patients with childhood-onset epilepsy. This review includes neither a discussion of adverse outcomes of AED treatment, nor seizure outcome after AED withdrawal or seizure outcome in adult-onset epilepsy, unless the latter is mentioned for comparison with childhood-onset epilepsy. A number of reviews have covered the less than recent literature on predicting seizure outcome in childhood-onset epilepsy [1–7].

Methods

This article reviews AED response in terms of seizure outcome reported in population-based studies of patients with childhood-onset epilepsy published between 1990 and 2009, unless otherwise indicated. We have included population-based studies from Western countries based on two criteria, unless otherwise noted. The first criteria is follow-up of 10 years or more and the second is publications that include exact risk estimation data for individual factors.

The following definitions have been used for description of seizure outcome. Seizure remission is defined as the proportion of patients entering at least 5-year seizure freedom on or off medication or terminal remission of 5 years or more seizure freedom at the end of follow-up [4,8], unless otherwise indicated. Poor seizure outcome (synonym: persistent seizures, difficult-to-treat, intractable, refractory or drug-resistant epilepsy) is defined either as failure to enter remission, as noted above, or as defined in the individual study. A task force of the International League Against Epilepsy (ILAE) recently defined drug-resistant epilepsy as failure of adequate trials of two tolerated, appropriately chosen and used AED schedules (whether as monotherapies or in combination) to achieve sustained seizure freedom [9].

Seizure outcome

Most long-term seizure outcome studies examine the proportion of patients ever entering at least 5-year remission on or off medication or terminal remission of 5 years (5YTR) or more at the end of follow-up [4,8]. A number of factors may influence the seizure outcome of children with epilepsy (Figure 1)[10,11].

Long-term population-based outcome studies in childhood-onset epilepsy commonly report approximately 70% of the patients to be in 5-year remission [12]. In a Japanese study of childhood-onset epilepsy [12], 63% were in remission after the mean follow-up of 19 years. Patients with epilepsy tend to achieve seizure freedom over the course of time and the remission rate is subsequently dependent on the duration of follow-up. After the first 5-year remission, the 5YTR was maintained with incident epilepsy in 57, 58 and 68% of the patients at 10, 20 and 30 years from the diagnosis, respectively [4,5]. The outcome in childhood-onset epilepsy is similar to that reported in the well-known Rochester study, which included adults and children. Annegers et al. reported the percentage of patients that demonstrate 5 consecutive seizure-free years within 10 years after diagnosis to be 65% and within 20 years to be 76% [13]. From the viewpoint of achieving remission, favorable predictors were idiopathic/cryptogenic versus symptomatic etiology (74 vs 46%); generalized tonic–clonic seizures versus focal seizures (85 vs 65%); and age of less than 10 years versus 10–19 years versus more than 19 years (85 vs 68 vs 63%, respectively).

Predicting long-term outcome

Table 1 summarizes predictors of epilepsy outcome reported in studies using multivariable analysis. Early liability to seizures may be manifested as neonatal seizures that are, with the exception of rare genetic epilepsies, caused by underlying structural abnormalities. Structural brain pathology may appear as neurodeficits (cerebral palsy or mental retardation/learning disability [LD]) or MRI lesions and can predispose patients to early onset of epilepsy. A high frequency of seizures before and during the early stage of treatment and seizure clustering during treatment predict difficult-to-treat or even intractable epilepsy [14–17].

Gender & age

In a long-term incident cohort of childhood-onset epilepsy, in the group who had weekly seizures during the first year of treatment (which indicated a poorer seizure outcome), two-thirds were males (66 vs 34%) [18]. In a study from The Netherlands that did not examine seizure frequency during the first year of treatment, no significant difference was found between boys and girls with regard to the following variables at study onset: sex, age, seizure type, type of epilepsy, etiology, pre-existing neurological signs, postictal signs, family history, standard EEG at intake and temporal seizure pattern [19]. However, when the abovementioned variables were combined at study onset and at 6 months, boys had a 1.64-fold higher risk for not achieving 1-year terminal remission within 5-year follow-up [19]. Whether early childhood age at onset is a predictor of seizure outcome is controversial [20].

Etiology

In a long-term cohort of childhood-onset epilepsy, patients with symptomatic etiology failed to enter 5YTR significantly more often than the idiopathic/cryptogenic group [8,15]. However, even the remote symptomatic group achieved cumulative 5-year remission at 9–10 years follow-up of more than 60%. An underlying congenital cause or early childhood brain damage causing neurodeficits is a strong predictor for a poor long-term seizure outcome [21]. Finally, there is compelling evidence from a community-based cohort of 77 children with new-onset temporal lobe epilepsy, who were followed prospectively with formal review at 7 and 14 years following seizure onset, that lesions on MRI, but not initial seizure frequency or early seizure remissions, were predictive of seizure outcome [22].

Neuropsychological comorbidity

Children with epilepsy and neurodeficits bear a less favorable outlook for long-term remission [21]. In a long-term followed population-based study of 242 patients with childhood-onset epilepsy [23], risk factors for any LD (whether associated with mental retardation or not) in multivariate analysis included occurrence of cerebral palsy, onset of epilepsy before the age of 6 years, and poor early effect of therapy. Among mentally normal or near normal (IQ>70) individuals, the only significant predictor for LD was a symptomatic etiology. Subjects with no LD had a 4.3-fold increased chance of 5-year remission compared with those with LD. Furthermore, the risk for relapse after a 5-year remission (with or without medication) was 3.2-fold in patients with LD compared with those who had no LD [23].

An overall incidence of psychiatric disorders of 14% was found in a long-term follow-up study of patients with childhood-onset epilepsy. Psychotic disorders were identified in 3% [24]. The most common psychoses were manic–depressive psychosis and paranoid schizophrenia. In a Canadian population-based study of intellectually normal children with epilepsy, 22% were assessed to need mental health consultation for behavioral disturbances. Two significant predictors for poor seizure outcome were found: LD and greater than 21 seizures prior to treatment [18]. Psychiatric comorbidity was not found to be a predictor of poor seizure outcome [18].

Post-traumatic brain injury

Brain injury may be a cause or consequence of epilepsy. In a retrospective population study of children up to 18 years of age, following a post-traumatic brain injury, the risk for future epilepsy was increased in line with the severity of the trauma [25], but even mild trauma increased the risk in children and young adults [26].

Status epilepticus

During 5-year follow-up, children with epilepsy, but without a history of status epilepticus achieved a >1-year remission (1YR) nonsignificantly less often than those without status epilepticus before intake or during follow-up [27]. However, in a long-term follow-up of childhood-onset epilepsy, status epilepticus was a predictor for not achieving remission [28]. The reasons for the discrepancy between the two studies are not clear, but there were some considerable methodological differences between the studies. In the Dutch study, one status epilepticus was enough to fulfill the definition of epilepsy [27], while the long-term study, followed the ILAE definition of epilepsy [28]. The follow-up period was also different between the studies at 5 years [27] compared with at least 30 years [28].

Seizure type & epilepsy syndrome

In a Finnish population-based long-term incident study of childhood-onset epilepsy [8], a remission of at least 5 years was achieved by 29 out of 45 (64%) people with generalized epilepsy. In the generalized epilepsy group, remission was found in 27 out of 31 (87%) patients with idiopathic epilepsy, 14 out of 15 (93%) individuals with generalized tonic–clonic seizures only (for a definition, see [7]), and in 13 out of 14 (93%) patients with other generalized epilepsies. In those with localization-related epilepsy, 57 out of 86 (66%) patients were in 5-year remission. Within the localization-related epilepsies, remission was found in 13 out of 14 (93%) patients with rolandic epilepsy, in 39 out of 65 (60%) with focal symptomatic epilepsy and in 23 out of 43 (53%) with temporal lobe epilepsy. Among patients with other localization-related epilepsy, 21 out of 29 (72%) became seizure free [8].

For a comparison, a mixed study including children and adults, demonstrated that, at 10 years of follow-up, there was a 5-year remission in 68–75% of cases with idiopathic generalized tonic–clonic seizures [13]. In childhood absence epilepsy, applying strict diagnostic criteria [29], 95% entered remission [30], while in juvenile absence epilepsy only 44% (8 out of 17) of seizures were controlled after the mean 6-year duration of follow-up [31]. Of patients with juvenile myoclonic epilepsy, 44–74% become seizure free on drug treatment [32–34].

Catastrophic epilepsies of childhood

Patients with West syndrome or Lennox–Gastaut syndrome (LGS) are most unlikely to become seizure free. In 20–30 years of follow-up, a 5-year remission can be expected in approximately a third of surviving patients [35,36]. The etiology of the syndrome is a strong predictor for cognitive outcome. Mental retardation occurs in 30–50% with cryptogenic etiology and in 80–95% with symptomatic etiology [37]. LGS is often (in 13–65% of cases), and in cases of symptomatic etiology in particular, a continuation of West syndrome [36,38,39]. However, as shown in a community-based study, prenatal or perinatal abnormalities are not necessarily in correlation with the severity of epilepsy [40]. At follow-up of at least 10 years, 5–13% of individuals were in remission and 8% were not mentally retarded. Patients with cryptogenic LGS had a better cognitive outcome compared with patients who had symptomatic LGS [38], but another study reported no difference between the groups [39].

Seizure frequency & seizure clustering

Prospective studies of outcome in populations of children and adults with newly treated epilepsy have consistently shown that the single most important factor associated with the chance of remission of seizures is the frequency of seizures in the early phase of epilepsy, with an association between increased number of seizures in this period and poorer outcome [8,41]. Two factors, present early in the course of treatment, were found to be associated with poor seizure outcome. Having weekly seizures during the first year of treatment carried a eightfold increased risk (hazard ratio [HR]: 8.2 [95% CI: 1.6–43.0]; p = 0.0125) of developing drug-resistant epilepsy and a twofold increased risk of never entering terminal 1YR (HR: 2.7 [95% CI: 1.5–5.0]; p = 0.0010). Having weekly seizures prior to treatment only slightly increased the risk of never entering terminal 1YR (HR: 1.7 [95% CI: 1.04–2.9]; p = 0.0350) [15]. To provide evidence as to whether seizure clustering is associated with drug resistance, a prospective, long-term population-based study was performed. In total, 120 of the 150 patients with childhood-onset epilepsy who had been followed since the onset of the disorder (for an average of 37.0 years, standard deviation: 7.1, median: 40.0; range: 11–42; incident, i.e., followed-up since the first seizure) [8]. At the end of follow-up, 26 out of 120 patients (22%) had recorded clusters of seizures. A total of 14 out of 26 patients had a cluster before drug therapy (10 out of 14 as first seizures) and 12 out of 26 during drug treatment. Seizure cluster occurring during drug therapy was significantly negatively associated with both 5YTR (p = 0.0039) and with 5-year remission (p = 0.0230). By contrast, patients with seizure clustering prior to treatment versus no clustering showed no difference in seizure outcome [14]. Despite symptomatic etiology, as indicated by lesional MRI, being shown in a hospital-based study to be a risk factor even for children with low seizure frequency [22], a population-based study has demonstrated that those with symptomatic etiology plus early weekly seizures have a poorer seizure outcome compared with those with symptomatic etiology and less frequent seizures [15]. Finally, it is important in that regard to note that despite having either a weekly seizure frequency or symptomatic epilepsy, 76–81% of children will have entered 1YR at 10 years follow-up. In addition, the proportion of children entering 1YR in both groups will increase with further long-term follow-up. Despite having either a weekly seizure frequency or symptomatic epilepsy, 63–66% of children will have entered 1-year terminal remission at 40 years of follow-up. Even in the presence of a combination of frequent seizures in the first year of treatment and symptomatic etiology, 19 out of 24 (79%) children will enter one or more periods of 1YR, and 1-year terminal remission will be achieved in nine out of 24 (37%) in the course of their disorder. However, none of these 24 children can expect to remain seizure-free from the start of treatment to the end of follow-up. It is reassuring that virtually all (51 out of 52 [98%]) children with low seizure frequency and nonsymptomatic etiology will enter 1YR during the 40 years of follow-up, and almost all (49 out of 52 [94%]) will enter 1-year terminal remission. In addition, as we have shown earlier, nearly one in five children whose epilepsy is initially well controlled will later develop drug resistance [8].

Prior AED treatment

Observing treatment outcome with AEDs provides insights into risk factors for seizure outcome. Failure of the first-ever AED treatment in newly diagnosed epilepsy is a risk factor for poor seizure outcome. Logistic regression indicated that the best model to predict refractory temporal lobe epilepsy contained only the variable ‘failure of first AED trial’, with a positive predictive value of 0.89 (95% CI: 0.76–0.96) and negative predictive value of 0.95 (95% CI: 0.87–0.99) to predict refractory temporal lobe epilepsy at 2 years [41]. This hospital-based, retrospective study of children shows that failure of first AED trial accurately predicts refractory temporal lobe epilepsy at 2 years after onset. If confirmed in a prospective setting and with longer follow-up, this finding should support earlier consideration of surgical options [41]. More recently, the same group developed a model to predict the outcome of treatment with carbamazepine in children with newly diagnosed focal epilepsy of presumed temporal lobe origin, using data available at the time of diagnosis [42]. A total of 149 patients completed an adequate first AED trial. Carbamazepine was the initial drug used in 129 (87%) patients. A total of 41 of these 129 patients (32%) had persistent seizures. Significant predictors of initial carbamazepine failure were as follows: early risk factor for epilepsy (risk ratio: 3.1 [95% CI: 1.6–4.0]) and temporal lobe abnormality on MRI scan (risk ratio: 3.1 [95% CI: 1.7–4.2]). The outcome of the initial carbamazepine trial was correctly classified in up to 78% of patients. Accurate prediction of initial carbamazepine failure was as high as the predictive value of 0.67 (95% CI: 0.53–0.79). Accurate prediction of initial carbamazepine success was as high as the predictive value of 0.87 (95% CI: 0.77–0.94). In this hospital-based study, standard clinical data were less than adequate for predicting response to the initial trial of carbamazepine, with prediction of carbamazepine failure being particularly difficult [43]. This study suggests that better markers of AED response and nonresponse are required to guide optimal therapy in patients with epilepsy. In a Dutch population-based study of childhood-onset epilepsy, a remission of at least 1 year at 5 years follow-up was attained by 46% of patients on the first AED, by 19% on the second AED, and by 9% on all additional AED regimens [19]. Almost 60% of the children treated with a second or additional AED regimen had a remission of at least 1 year, showing that after failure of the first AED, treatment can still be successful [19]. In a prospective population-based study of childhood-onset epilepsy, delayed time to first remission identified poor long-term drug response of childhood-onset epilepsy [44]. The authors assessed time to first 1YR as a determinant of entering future 5YTR in a population-based cohort of 144 children prospectively followed up since their first unprovoked seizure before the age of 16 years up to the mean age of 48 years. The proportion of patients entering 5YTR was highly dependent on the length of time to first 1YR after starting adequate treatment. For 144 patients, the overall 5YTR rate decreased from 32% for those in remission at year 1 to 24% at year 2, to 5% after 3 and 4 years, to 2% after 5 years or longer. Patients who entered 1YR within the first 5 years of treatment had an 11-fold improved chance of entering 5YTR (odds ratio = 11.4; 95% CI: 2.9–45.3; p = 0.0005) and a ninefold chance for uninterrupted 5YTR off medications (odds ratio: 9.0; 95% CI: 1.17–1.9; p = 0.0383) compared with those who did not enter 1YR within the first 5 years of treatment. Three additional independent prognostic factors for predicting 5YTR were confirmed: etiology, seizure frequency prior to treatment and seizure frequency during treatment. The study authors concluded that delayed efficacy after starting drug treatment gradually diminishes chances for long-term seizure remission, whether on medication or not. Not entering remission within 5 years of starting treatment predicts failure to achieve long-term seizure freedom in the future for the vast majority of patients [44].

Methodological concerns

The variability and the range of factors predicting seizure outcome of treatment can be explained at least in part by a number of pitfalls in the design of the published reports. Most of the studies dealing with factors predicting long-term seizure outcome are observational, retrospective, short-lasting, clinic-based and carried out in small samples of patients. Lack of specificity and sensitivity of observed risk factors is of concern [45]. Lack of predefined definitions of seizure-free periods, short duration of follow-up, small sample size and selection bias all contribute to a biased assessment of the treatment outcome and to a biased identification of the risk factors. Furthermore, predicting long terminal remission has been shown to be more successful than predicting continuing seizures [19]. In the absence of a randomized control group without treatment, one cannot assess whether the observed outcome reflects changes in the natural history of epilepsy. In addition, risk factors for seizure outcome should be predefined and, if possible, assessed without knowledge of the patient’s status (whether treated or untreated).

Expert commentary

Based on the available evidence, which includes studies with a small sample size, our review indicates that the long-term seizure outcome of a child with epilepsy can be predicted in many cases at the very first visit and in all cases during the first years of treatment. Although remote symptomatic etiology of the epilepsy as shown by MRI or neurocognitive deficits are risk factors for drug-resistant epilepsy, up to 60% of new-onset patients will enter remission with early treatment. Children with idiopathic epilepsy and those without MRI or neurocognitive deficits and no history of status epilepticus usually have treatable epilepsy resulting in early and long-lasting remission. The take-home messages from this article are first, that the outcome of childhood-onset epilepsy can be predicted early in most cases; second, that commonly accessible diagnostic features, such as remote symptomatic etiology, abnormal intelligence and having one of the symptomatic generalized epilepsy syndromes, generally allow the prediction of a poorer prognosis for childhood-onset epilepsy; third, the fact outlined previously that a small majority of cases do enter remission despite the presence of these poor prognostic factors shows that our current ability to predict seizure outcome is limited. The obvious conclusion, which raises concern, is that we are missing important factors that determine or at least allow us to predict seizure outcome in childhood-onset epilepsy. The clinical features that were examined for their prognostic validity in the past are useful, as outlined in this review, but help us to recognize that they are just the ‘tip of the iceberg’. Equally important and perhaps even more important features that determine the natural history of the child’s epilepsy remain elusive to clinical examination at present. Candidate factors to better predict the prognosis of childhood-onset epilepsy are neurobiological factors that determine the individual’s epileptogenesis prior to the first seizure and, perhaps, even during the early years of epilepsy.

Animal models of epilepsy and human tissue studies suggest that epileptogenesis involves a cascade of molecular, cellular and neuronal network alterations [46]. Within minutes to days following the initial insult, there are acute early changes in neuronal networks, which include rapid alterations to ion channel kinetics as a result of membrane depolarization, post-translational modifications to existing functional proteins and activation of immediate early genes. Subacute changes occur from hours to weeks and include transcriptional events, neuronal death and activation of inflammatory cascades. The chronic changes that follow over weeks to months include anatomical changes, such as neurogenesis, mossy fiber sprouting, network reorganization and gliosis. These epileptogenic processes are developmentally regulated and might contribute to differences in epileptogenesis between adult and developing brains and, perhaps, between drug-responsive and drug-resistant cases of epilepsy. An understanding of these factors could yield potential biomarkers for the prevention of epileptogenesis and, more importantly for the topic of this article, also provide biomarkers for identifying patients at risk of developing drug-resistant epilepsy.

Five-year view

Although knowledge regarding the predictive factors for seizure outcome of children with epilepsy is informative, the challenge for the next 5 years is to translate this knowledge into clinical benefit. To achieve higher rates of seizure freedom, this may mean being able to eliminate symptomatic causes of epilepsy such as prenatal or postnatal brain damage through injury and status epilepticus. Better perinatal care may decidedly improve the occurrence rate of epilepsy in childhood. Another area for improvement of seizure outcome in children is to tackle the elusive causes of mental retardation. More aggressive treatment early in the course of epilepsy through modern drugs may avoid children entering into the dark universe of drug-resistant epilepsy with high morbidity. One further challenge is to identify factors that predict epileptogenesis – that is, the underlying disease process that entertains the occurrence of seizures as symptoms. The quest for finding a cure for epilepsy includes a search for predictive factors that drive the development from injury of the brain in the widest sense to the development of the first seizure and, furthermore, what propels the disease to become intractable to drugs.

Table 1. Reported factors predicting drug response in multivariate analyses of individual population-based studies in Western countries that reported a follow-up of 10 years or more.

Variable	Number in remission	Risk estimation (95%CI)	p-value	Study authors (total number of cases; average follow-up in years)	Ref.
Predictors of seizure remission on or off AEDs
Pre-seizure factors
Remote symptomatic etiology	37/83 (45%) vs 46/83 (55%)	RR: 0.8 (0.6–1.1)	ns overall cohort	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
	35/74 (47%) vs 39/74 (53%)	RR: 0.9 (0.7–1.2)	ns overall incident cohort^†	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
Normal vs abnormal intelligence	100/130 (77%) vs 24/64(38%)	RR: 2.1 (1.5–2.9)	p < 0.001	Brorson and Wranne (1987) (n = 194; 12 years)	[3]
Normal vs abnormal neurological examination	108/151 (72%) vs 16/43(35%)	RR: 1.9 (1.3–2.9)	p < 0.001	Brorson and Wranne (1987) (n = 194; 12 years)	[3]
Seizure frequency 2 per 6 months vs >2 per 6 months	85/109 (78%) vs 39/46(46%)	RR: 1.7 (1.3–2.2)	p < 0.001	Brorson and Wranne (1987) (n = 194; 12 years)	[3]
One seizure type vs several seizure types	89/123 (72%) vs 35/71(49%)	RR: 1.5 (1.1–1.9)	p = 0.001	Brorson and Wranne (1987) (n = 194; 12 years)	[3]
Localization-related idiopathic vs others	13/14 (93%) vs 84/130 (65%)	RR: 1.4 (1.2–1.7)	p = 0.0361	Sillanpää and Schmidt (2006) (n = 144; 37 years)	[8]
Localization-related symptomatic vs others	39/65 (0%) vs 58/79 (73%)	RR: 0.8 (0.6–1.0)	p = 0.00875	Sillanpää and Schmidt (2006) (n = 144; 37 years)	[8]
Temporal lobe vs others	23/43 (53%) vs 74/101(73%)	RR: 0.7 (0.5–1.0)	p = 0.0205	Sillanpää and Schmidt (2006) (n = 144; 37 years)	[8]
Generalized vs others	29/45 (64%) vs 68/99 (69%)	RR: 0.9 (0.7–1.2)	p = 0.6148	Sillanpää and Schmidt (2006) (n = 144; 37yrs)	[8]
Generalized idiopathic vs others	27/31 (87%) vs 70/113 (62%)	RR: 1.4 (1.2–1.7)	p = 0.0082	Sillanpää and Schmidt (2006) (n = 144; 37 years)	[8]
Generalized cryptogenic/symptomatic vs others	2/14 (14%) vs 95/130 (73%)	RR: 0.2 (0.1–0.7)	p < 0.0001	Sillanpää and Schmidt (2006) (n = 144; 37 years)	[8]
Seizure-related factors
Complex partial seizures	35/74 (47%) vs 39/74 (53%)	RR: 0.3 (0.2–0.5)	p < 0.001 overall cohort	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
Complex partial seizures	15/43 (35%) vs 39/74 (53%)	RR: 03 (0.1–0.6)	p 0.001 overall incident cohort^†	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
Atonic seizures	3/14 (21%) vs 11/14 (79%)	RR: 0.2 (0.1–0.7)	p < 0.002 overall cohort	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
Atonic seizures	35/74 (47%) vs 39/74 (53%)	RR: 0.3 (0.2–0.5)	p < 0.001 overall incident cohort^†	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
Treatment-related factors
Early response to therapy	80/102 (78%) vs 40/102(39%)	RR: 3.7 (2.0–6.7)	p < 0.001 overall cohort	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
Early response to therapy	50/73 (68%) vs 23/73(32%)	RR: 2.2 (1.2–4.0)	p < 0.001 incident cohort	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
Entering 1 year remission within first 5 years of treatment	93/97 (96%) vs 27/47 (57%)	RR: 4.7 (1.9–11.4)	p < 0.001	Sillanpää and Schmidt (2006) (n = 144; 37 years)	[8]
Likelihood for 5YTR off AEDs	45/120(45%) vs 1/8(8%)	OR: 9.0 (1.1–71.9)	p = 0.0383	Sillanpää and Schmidt (2009)	[44]
Predictors of not entering seizure remission
Pre-seizure factors
Symptomatic etiology	37/83 (33%) vs 75/93 (81%)	RR: 0.6 (0.4–0.7)	p < 0.001	Sillanpää et al. (1998) (n = 176; 31 years)	[2]
Normal vs abnormal intelligence	23/74(31%) vs 39/74(53%)	RR: 0.6 (0.4–0.9)	p < 0.05	Brorson and Wranne (1987) (n = 194; 12 years)	[3]
Seizure-related factors
Weekly vs less than weekly seizures prior to treatment	23/37 (62%) vs 55/65 (85%)	HR: 1.7 (1.04–2.9)	p = 0.00350 for never entering 1-year remission	Sillanpää and Schmidt (2009) (n = 102; 37 years)	[15]
Frequent seizures (undefined)	46/85 (54%) vs 24/109 (22%)	RR: 2.4 (1.6–3.7)	p < 0.01	Brorson and Wranne (1987) (n = 194; 12 years)	[3]
Several types of seizures	36/71(51%) vs 34/123 (28%)	RR: 1.8 (1.3–2.6)	p = 0.001	Brorson and Wranne, (1987) (n = 194; 12 years)	[3]
Seizure-cluster vs no seizure cluster prior to drug treatment	5/12(42%) vs 12/94 (13%)	RR: 3.3 (1.4–7.7)	p = 0.012	Sillanpää and Schmidt (2008) (120; 37 years)	[14]
Treatment-related factors
Weekly vs less than weekly seizures during first year of treatment	2/8 (25%) vs 18/19 (95%)	HR: 8.2 (1.6–43.0)	p = 0.0125 for drug-resistant epilepsy	Sillanpää and Schmidt (2009) (n = 102; 37 years)	[15]
Weekly vs less than weekly seizures during first year of treatment	During first year of treatment 19/38 (50%) vs 59/64 (92%)	HR: 2.7 (1.5–5.0)	p = 0.001 never entering terminal 1-year remission	Sillanpää and Schmidt (2009) (n = 102; 37 years)	[15]

^†Remission off medication.

Predictors of entering 5-year remission and predictors of not entering remission, each as defined by the study authors, are given separately.

5YTR: 5-year terminal remission; AED: Antiepileptic drug; HR: Hazard ratio; ns: Not significant; OR: Odds ratio; RR: Rate ratio.

Key issues

• • Based on the available evidence, which includes studies with a small sample size, our article indicates that long-term seizure outcome of a child with epilepsy can be predicted in many cases at the very first visit and in all cases during the first years of treatment.

• • Although remote symptomatic etiology of the epilepsy as shown by MRI or neurocognitive deficits are risk factors for drug-resistant epilepsy, up to 60% of new-onset patients will enter remission with early treatment.

• • Children with idiopathic epilepsy and those without MRI or neurocognitive deficits and no history of status epilepticus usually have well-treatable epilepsy, resulting in early and long-lasting remission.

• • A small majority of cases do enter remission despite the presence of these poor prognostic factors. This shows that our current ability to predict seizure outcome is limited.

• • The obvious conclusion, which raises concern, is that we are missing important factors that determine or at least allow us to predict seizure outcome in childhood-onset epilepsy.

• • Candidate factors to better predict the prognosis of childhood-onset epilepsy are neurobiological factors, including molecular, cellular and neuronal network alterations, that determine the individual’s epileptogenesis – that is, the process of developing epilepsy – prior to the first seizure and, perhaps, even during the early years of epilepsy.

• • The chronic changes that follow in the weeks to months after a first seizure include anatomical changes, such as neurogenesis, mossy fiber sprouting, network reorganization and gliosis. These epileptogenic processes are developmentally regulated and might contribute to differences in epileptogenesis between adult and developing brains, and, perhaps, between drug-responsive and drug-resistant cases of epilepsy.

• • An understanding of these factors could yield potential biomarkers for identifying patients at risk of developing drug-resistant epilepsy.

References

• Sillanpää M. Social functioning and seizure status of young adults with onset of epilepsy in childhood. An epidemiological 20-year follow-up study. Acta Neurol. Scand.96(Suppl.), 1–81 (1983).
   Google Scholar
• Sillanpää M, Jalava M, Kaleva O, Shinnar S. Long-term prognosis of seizures with onset in childhood. N. Engl. J. Med.338, 1715–1722 (1998).
   PubMed Web of Science ®Google Scholar
• Brorson LO, Wranne L. Long-term prognosis in childhood epilepsy: survival and seizure prognosis. Epilepsia28, 324–330 (1987).
   Google Scholar
• Sillanpää M. Long-term outcome of epilepsy. Epileptic Disord.2, 79–88 (2000).
   PubMedGoogle Scholar
• Sillanpää M. Children with epilepsy as adults: outcome after 30 years of follow-up. Acta Paediatr. Scand. Suppl.368, 1–78 (1990).
   PubMedGoogle Scholar
• Okuma T, Kumashiro H. Natural history and prognosis of epilepsy: report of a multi- institutional study in Japan. The group for the study of prognosis of epilepsy in Japan. Epilepsia22, 35–53 (1981).
   PubMed Web of Science ®Google Scholar
• Janz D. [The Epilepsies]. Thieme, Stuttgart, Germany (1969).
   Google Scholar
• Sillanpää M, Schmidt D. Natural history of treated childhood-onset epilepsy: prospective, long-term population-based study. Brain129, 617–624 (2006).
   PubMed Web of Science ®Google Scholar
• Kwan P, Arzimanoglou A, Berg AT et al. Definition of drug resistant epilepsy: Consensus proposal by the ad hoc Task Force of the ILAE Commission on Therapeutic Strategies. Epilepsia51(6), 1069–1077 (2009).
   PubMed Web of Science ®Google Scholar
• Sillanpää M, Camfield P, Camfield C. Predicting long-term outcome of childhood epilepsy in Nova Scotia, Canada, and Turku, Finland. Validation of a simple scoring system. Arch. Neurol.52, 589–592 (1995).
   Google Scholar
• Ko TS, Holmes GL. EEG and clinical predictors of medically intractable childhood epilepsy. Clin. Neurophysiol.110, 1245–1251 (1999).
   PubMed Web of Science ®Google Scholar
• Wakamoto H, Nagao H, Hayashi M, Morimoto T. Long-term medical, educational, and social prognoses of childhood-onset epilepsy: a population-based study in a rural district of Japan. Brain Dev.22(4), 246–255 (2000).
   PubMed Web of Science ®Google Scholar
• Annegers JF, Hauser WA, Elveback LR. Remission of seizures and relapse in patients with epilepsy. Epilepsia20(6), 729–737 (1979).
   PubMed Web of Science ®Google Scholar
• Sillanpää M, Schmidt D. Seizure clustering during drug treatment affects seizure outcome and mortality of childhood-onset epilepsy. Brain131, 938–944 (2008).
   PubMed Web of Science ®Google Scholar
• Sillanpää M, Schmidt D. Early seizure frequency and aetiology predict long-term medical outcome in childhood-onset epilepsy. Brain132, 989–998 (2009).
   PubMed Web of Science ®Google Scholar
• Beghi E, Tognoni G. Prognosis of epilepsy in newly referred patients: a multicenter prospective study. Collaborative Group for the Study of Epilepsy. Epilepsia29(3), 236–243 (1988).
   PubMedGoogle Scholar
• Kwan P, Brodie MJ. Early identification of refractory epilepsy. N. Engl. J. Med.342, 314–319 (2000).
   PubMed Web of Science ®Google Scholar
• Camfield C, Camfield P, Smith B, Gordon K, Dolley J. Biological factors as predictors of social outcome of epilepsy in intellectually normal children: a population-based study. J. Pediatr.122, 869–873 (1993).
   Google Scholar
• Arts WF, Brouwer OF, Peters AC et al. Course and prognosis of childhood epilepsy: 5-year follow-up of the Dutch study of epilepsy in childhood. Brain127, 1774–1784 (2004).
   PubMed Web of Science ®Google Scholar
• Camfield C, Camfield P, Gordon K, Smith B, Dooley J. Outcome of childhood epilepsy: a population-based study with a simple predictive scoring system for those treated with medication. J. Pediatr.122, 861–868 (1993).
   PubMed Web of Science ®Google Scholar
• Sillanpää M. Remission of seizures and predictors of intractability in long-term follow-up. Epilepsia34, 930–936 (1993).
   PubMed Web of Science ®Google Scholar
• Spooner CG, Berkovic SF, Mitchell LA, Wrennall JA, Harvey AS. New-onset temporal lobe epilepsy in children: lesion on MRI predicts poor seizure outcome. Neurology67, 2117–2118 (2006).
   Google Scholar
• Sillanpää M. Learning disability: occurrence and long-term consequences in childhood-onset epilepsy. Epil. Behav.5, 937–944 (2004).
   PubMed Web of Science ®Google Scholar
• Jalava M, Sillanpää M. Concurrent illnesses in adults with childhood-onset epilepsy: a population-based 35-year follow-up study. Epilepsia37, 1155–1163 (1996).
   Google Scholar
• Emanuelsson I, Uvebrant P. Occurrence of epilepsy during the first 10 years after traumatic brain injury acquired in childhood up to 18 years in the south western Swedish population-based series. Brain Inj.23, 612–616 (2009).
   Google Scholar
• Christensen J, Pedersen MG, Sidenius P, Olsen J, Vestergaard M. Long-term risk of epilepsy after traumatic brain injury in children and young adults: a population-based cohort study. Lancet373, 1105–1110 (2009).
   PubMed Web of Science ®Google Scholar
• Stroink H, Geerts AT, van Donselaar CA et al. Status epilepticus in children with epilepsy: Dutch study of epilepsy in childhood. Epilepsia48, 1708–1715 (2007).
   PubMedGoogle Scholar
• Sillanpää M, Shinnar S. Status epilepticus in a population-based cohort with childhood-onset epilepsy in Finland. Ann. Neurol.52, 303–310 (2002).
   PubMedGoogle Scholar
• Panayiotopoulos CP. Absence epilepsies. In: Epilepsy: A Comprehensive Textbook. Engel J Jr, Pedley TA (Eds). Lippincott-Raven, PA, USA, 2327–2346 (1997).
   Google Scholar
• Grosso S, Galimberti D, Vezzosi P et al. Childhood absence epilepsy: evolution and prognostic factors. Epilepsia46, 1796–2801 (2005).
   Google Scholar
• Tovia E, Goldberg-Stern H, Shahar E, Kramer U. Outcome of children with juvenile absence epilepsy. J. Child Neurol.21, 766–768 (2006).
   Google Scholar
• Sharpe C, Buchanan N. Juvenile myoclonic epilepsy: diagnosis, management and outcome. Med. J. Aust.162, 133–134 (1995).
   PubMedGoogle Scholar
• Martínez-Juárez IE, Alonso ME, Medina MT et al. Juvenile myoclonic epilepsy subsyndromes: family studies and long-term follow-up. Brain129, 1269–1280 (2006).
   PubMed Web of Science ®Google Scholar
• Camfield CS, Camfield PR. Juvenile myoclonic epilepsy 25 after seizure onset: a population-based study. Neurology73, 1041–1045 (2009).
   PubMed Web of Science ®Google Scholar
• Riikonen R. Long-term outcome of West syndrome: a study of adults with a history of infantile spasm. Epilepsia37: 367–372. Epilepsia37, 367–372 (1996).
   PubMed Web of Science ®Google Scholar
• Camfield P, Camfield C. Long-term prognosis for symptomatic (secondarily) generalized epilepsies: a population-based study. Epilepsia48, 1128–1132 (2007).
   PubMed Web of Science ®Google Scholar
• Wong M, Trevathan E. Infantile spasms. Ped. Neurol.24, 89–98 (2001).
   Google Scholar
• Goldsmith IL, Zupanc ML, Buchhalter JR. Long-term seizure outcome in 74 patients with Lennox–Gastaut syndrome: effects of incorporating MRI head imaging in defining the cryptogenic subgroup. Epilepsia41, 395–399 (2000).
   PubMed Web of Science ®Google Scholar
• Rantala H, Putkonen T. Occurrence, outcome, and prognostic factors of infantile spasms and Lennox–Gastaut syndrome. Epilepsia40, 286–289 (1999).
   Google Scholar
• Heiskala H. Community-based study of Lennox–Gastaut syndrome. Epilepsia38, 526–531 (1997).
   PubMed Web of Science ®Google Scholar
• Hitiris N, Mohanraj R, Norrie J, Sills GJ, Brodie MJ. Predictors of pharmacoresistant epilepsy. Epilepsia Res75, 192–196 (2007).
   Google Scholar
• Dlugos DJ, Sammel MD, Strom BL, Farrar JT. Response to first drug trial predicts outcome in childhood temporal lobe epilepsy. Neurology57, 2259–2264 (2001).
   PubMed Web of Science ®Google Scholar
• Dlugos DJ, Buono RJ. Predicting outcome of initial treatment with carbamazepine in childhood focal epilepsy. Pediatr. Neurol.30, 311–315 (2004).
   Google Scholar
• Sillanpää M, Schmidt D. Delayed time to first remission identifies poor long-term drug response of childhood-onset epilepsy: a prospective population-based study. Epil. Behav.16, 507–511 (2009).
   PubMed Web of Science ®Google Scholar
• Geerts AT, Arts WF, Brouwer OF et al. Validation of two prognostic models predicting outcome at two years after diagnosis in a new cohort of children with epilepsy: the Dutch Study of Epilepsy in Childhood. Epilepsia47, 960–965 (2006).
   Google Scholar
• Rakhade SN, Jensen FE. Epileptogenesis in the immature brain: emerging mechanisms. Nat. Rev. Neurol.5, 380–391 (2009).
   PubMed Web of Science ®Google Scholar

Predicting antiepileptic drug response in children with epilepsy

To obtain credit, you should first read the journal article. After reading the article, you should be able to answer the following, related, multiple-choice questions. To complete the questions and earn continuing medical education (CME) credit, please go to http://www.medscape.org/journal/expertneurothera. Credit cannot be obtained for tests completed on paper, although you may use the worksheet below to keep a record of your answers. You must be a registered user on Medscape.org. If you are not registered on Medscape.org, please click on the New Users: Free Registration link on the left hand side of the website to register. Only one answer is correct for each question. Once you successfully answer all post-test questions you will be able to view and/or print your certificate. For questions regarding the content of this activity, contact the accredited provider, [email protected]. For technical assistance, contact [email protected]. American Medical Association’s Physician’s Recognition Award (AMA PRA) credits are accepted in the US as evidence of participation in CME activities. For further information on this award, please refer to http://www.ama-assn.org/ama/pub/category/2922.html. The AMA has determined that physicians not licensed in the US who participate in this CME activity are eligible for AMA PRA Category 1 Credits™. Through agreements that the AMA has made with agencies in some countries, AMA PRA credit is acceptable as evidence of participation in CME activities. If you are not licensed in the US and want to obtain an AMA PRA CME credit, please complete the questions online, print the certificate and present it to your national medical association.

Activity Evaluation: Where 1 is strongly disagree and 5 is strongly agree

	1	2	3	4	5
1. The activity supported the learning objectives.
2. The material was organized clearly for learning to occur.
3. The content learned from this activity will impact my practice.
4. The activity was presented objectively and free of commercial bias.

1. You are asked to see a 3-year-old boy with seizures. Based on the above review by Drs. Sillanpää and Schmidt, which of the following statements about factors predating epilepsy that predict seizure remission is most likely correct?

• □ A Intelligence is not a factor predicting remission

• □ B Boys have a better prognosis for remission than girls

• □ C Neurologic examination is not a factor predicting remission

• □ D Remote symptomatic etiology based on MRI or neurocognitive deficits may increase risk of not achieving remission

2. Based on the above review, which of the following seizure-related factors in the patient described in question 1 is most likely to predict seizure remission in response to antiepileptic drug (AED) treatment?

• □ A Generalized idiopathic seizures

• □ B 4 seizures in a 6-month period

• □ C Temporal lobe seizures

• □ D Atonic seizures

3. Based on the above review, which of the following statements about treatment-related and other factors that predict seizure remission in patients with childhood-onset epilepsy treated with AEDs is most likely correct?

• □ A Early response to therapy is not a predictor of remission

• □ B Entering 1-year remission, on or off AEDs, within the first 5 years of treatment is not a predictor of maintaining remission

• □ C Children with poor prognostic factors inevitably require surgery

• □ D Neurobiologic factors, including molecular, cellular and neuronal network changes that determine epileptogenesis, may help predict prognosis of childhood-onset epilepsy