- Top of page
Purpose: To determine the prevalence and identify predictors of medical intractability in children presenting with epilepsy before 36 months of age, and to assess the effect of medical intractability on long-term mortality and intellectual function.
Methods: Children with newly diagnosed epilepsy before 36 months between 1980 and 2009 while resident in Olmsted County, MN, were identified. Medical records were reviewed to collect epilepsy-specific variables and long-term outcome data. Medically intractable epilepsy was defined as either (1) seizure frequency greater than every 6 months at final follow-up and failure of two or more antiepileptic drugs for lack of efficacy, or (2) having undergone epilepsy surgery after failure to respond to two or more antiepileptic drugs.
Key Findings: One hundred twenty-seven children with new-onset epilepsy were identified and followed for a median of 78 months. Medically intractable seizures occurred in 35%, and significant predictors on multivariate analysis were age ≤12 months at diagnosis (odds ratio [OR] 6.76, 95% confidence interval [CI] 2.00, 22.84, p = 0.002), developmental delay at initial diagnosis of epilepsy (OR 20.03, 95% CI 3.49, 114.83, p = 0.0008), neuroimaging abnormality (OR 6.48, 95% CI 1.96, 21.40, p = 0.002), and focal slowing on initial EEG (OR 5.33, 95% CI 1.14, 24.88, p = 0.03). Medical intractability occurred early in the course in most children, being seen in 61% by 1 year, and 93% by 5 years after initial diagnosis. Mortality was higher (20% vs. 0%, p < 0.001) and intellectual outcome poorer (p < 0.001) if epilepsy was medically intractable.
Significance: One third of children presenting with epilepsy before 36 months will be medically intractable, and significant predictors are identified. Medically intractable epilepsy is associated with increased mortality risk and significant intellectual disability.
Epilepsy is one of the most common neurologic disorders in children and has its highest incidence in the younger ages, with rates of 102.4 per 100,000 per year in the first year of life (Camfield et al., 1996; Wirrell et al., 2011).
Although population-based studies in children with epilepsy have shown a relatively favorable long-term outcome, with nearly two thirds achieving seizure freedom and nearly one half being able to discontinue antiepileptic drugs (Brorson & Wranne, 1987; Camfield & Camfield, 2003; Sillanpaa & Schmidt, 2006), no population-based study has specifically addressed outcomes in very young children. Epilepsy in early childhood may have a less favorable outcome for a number of reasons. Electroclinical syndrome remains one of the most robust predictors of outcome (Berg et al., 2001), and many severe epileptic encephalopathies, including West, Dravet, and Lennox-Gastaut syndromes, present early in life. Furthermore, specific etiologies including cortical dysplasias and other structural abnormalities (Semah et al., 1998), neurometabolic disorders, and perinatal brain injury are highly correlated with intractability and usually present in early childhood.
Early onset epilepsy is frequently associated with significant comorbidity. Intellectual disability frequently coexists with epilepsy in young children as a result of the underlying symptomatic cause of the seizures. However, there is also evidence that further decline in intellectual function may result from the sequelae of poorly controlled seizures in the developing brain (Berg et al., 2004b), emphasizing the importance of early recognition and effective therapy of intractable epilepsy in early childhood. Furthermore, the underlying structural and metabolic brain disturbances that lead to epilepsy, as well as ongoing uncontrolled seizures may result in greater morbidity and mortality due to profound neurodevelopmental delay. In a study that focused only on those with symptomatic generalized epilepsy, nearly one fourth had died, and most survivors had mental retardation and were highly dependent (Camfield & Camfield, 2007, 2008).
The goals of this study were to determine the prevalence and predictors of medical intractability in children with onset of epilepsy before 36 months of age, and to assess long-term mortality and intellectual function in those who met criteria for medical intractability.
- Top of page
In our population-based cohort, more than one third of children with onset of epilepsy in the first 3 years of life were medically intractable. This proportion is higher than in previously published population-based studies in new-onset childhood epilepsy, and is likely reflective of specific etiologies being more prevalent in very young children. In our cohort, we found structural causes to be extremely common, with malformations of cortical development and tuberous sclerosis accounting for nearly half, and other structural abnormalities (predominantly remote brain injury) accounting for one third of all medically intractable cases. Genetic syndromes accounted for only 7% of cases, and only 11% were of unknown etiology. Although all “unknown” etiology cases had undergone MRI, the majority presented with focal-onset seizures, and may have had less extensive forms of cortical dysplasia that are not well seen on MRI. Furthermore, the ability of MRI to detect cortical dysplasia can be limited in children between 6 and 24 months, due to evolution of myelination (Colombo et al., 2003). Specific epilepsy syndromes presenting early in life are known to be highly correlated with intractable epilepsy. In our cohort, a specific syndrome was identifiable at diagnosis in just over one fourth of medically intractable cases, with the majority being West syndrome. These syndromes are individually rare, and the number of affected patients is too low to reach statistical significance in most population-based studies.
Not all children with early onset epilepsy will do poorly, however. We found four significant predictors of medical intractability on multivariate analysis: age ≤12 months at diagnosis, developmental delay at initial diagnosis, abnormal neuroimaging, and focal slowing on initial EEG. These predictors, which are identifiable at the time of initial diagnosis, allowed more accurate prognostication of seizure outcome in our cohort presenting with epilepsy before 3 years of age. The association with medical intractability increased with the presence of each additional variable. In the absence of any variable, the probability of intractability was 0.01, whereas in the presence of all four variables, that probability rose to 0.97.
Our results are consistent with earlier work that has noted poorer outcomes in children with neuroimaging abnormalities, early onset seizures, and concomitant neurologic handicaps (Brorson & Wranne, 1987; Sillanpaa et al., 1995; Semah et al., 1998; Ko & Holmes, 1999; Camfield & Camfield, 2003). These findings support the concept that congenital or early acquired brain lesions frequently result in intractable epilepsy with significant impact on neurologic and cognitive function and suggest early surgical intervention should be considered for such cases.
Few studies have examined the predictive value of initial EEG on medical intractability in children. Similar to our results, Berg et al. (2001) found that focal slowing was predictive for intractability, however, in her study, interictal epileptiform discharges were not. Focal slowing may be seen postictally or represent an underlying cerebral dysfunction due to the epileptogenic lesion. Arts et al. (1999) reported that EEG abnormalities at intake were not significant, although such abnormalities at 6 months were predictive of poor outcome. In contrast, we found the presence of epileptiform discharges on initial EEG to be a significant risk factor for intractability. Furthermore, the type of epileptiform discharge was important—children with multifocal discharges or hypsarrhythmia were significantly more likely to develop intractable epilepsy than those with focal, hemispheric, or generalized discharge. Prior studies have noted that multifocal spikes are commonly seen in children with hypsarrhythmia on previous or follow-up EEG studies and that etiologies associated with multifocal spikes included the majority of diseases that afflict the brain early in life, often in a diffuse manner (Noriega-Sanchez & Markand, 1976; Blume, 1978; Kotagal, 1995).
Previous work has suggested that a substantial proportion of focal epilepsy may not become clearly intractable for many years after onset (Berg et al., 2003). Indeed, Berg et al. (2006) reported that 40% of children with focal epilepsy presented with delayed intractability (defined as >3 years after epilepsy onset). We found delayed intractability to be much less of a problem in this very young cohort. Medical intractability could be defined early in most children; 61% met criteria for intractability by 1 year and 86% by 3 years after initial diagnosis of epilepsy. In children with focal onset epilepsy, nearly two thirds were intractable by 1 year after onset, and nearly all (91%) met criteria by 3 years after onset.
Outcome in young children with medically intractable epilepsy was concerning. Twenty percent of our cohort had died after long-term follow-up, the majority due to conditions unrelated to their epilepsy, but as a result of their severe neurologic disability. Previous population-based or community-based cohort studies have shown mortality rates of 2.1–3.8% after long-term follow-up, and similar to our results, found that most deaths were unrelated to epilepsy (Camfield et al., 2002; Berg et al., 2004a). These studies have identified functional neurologic deficit, symptomatic etiology, and epileptic encephalopathy as predictors of higher mortality.
Because our study was retrospective, no consistent neuropsychological battery was regularly used, which limits a more detailed assessment of cognitive function. However, based on our best clinical assessment, we found intellectual disability to be prominent in our cohort of children with early onset medically intractable epilepsy, with only 5% having normal development, and nearly two thirds showing evidence of severe delay at final follow-up. Underlying structural abnormalities of the brain likely explain much of this disability. However, ongoing intractable seizures, very frequent epileptiform discharge and high doses of multiple medications contribute to epileptic encephalopathy and adversely affect development. Children who underwent epilepsy surgery showed lesser developmental disability after long-term follow-up, presumably due to greater likelihood of achieving seizure control, reduction of frequent epileptiform discharges on EEG, and reduced exposure to high-dose or multiple antiepileptic medications.
In summary, the results of our population-based, retrospective chart review show that just over one third of children with new-onset epilepsy before 36 months of age will develop medical intractability. Using the four main predictors of seizure outcome identified on multivariate analysis (age ≤12 months at onset, developmental delay at initial diagnosis, neuroimaging abnormalities, and focal slowing on initial EEG), outlined in Table 3, seizure outcome can be predicted with reasonable accuracy in the majority of young children with new-onset epilepsy before 36 months. Children with greater numbers of predictors should be monitored carefully, as intractability usually presents early in the course. Such children should be considered promptly for epilepsy surgery, if they are candidates, as medical intractability is associated with high mortality and significant intellectual disability.