Abstract
Anne T. Berg, Francine M. Testa, Susan R. Levy. Ann Neurol 2011;70:566–573.
OBJECTIVE: Determine the probability of attaining complete remission in children with nonsyndromic epilepsy (NSE) over the course of ≥10 years from initial diagnosis; identify early predictors of complete remission; and assess the risk of relapse after achieving complete remission. METHODS: In a prospective community-based cohort, complete remission was defined as 5 years seizure-free and medication-free. Any subsequent seizure for any reason was a relapse. Univariate and bivariate analyses were conducted with standard methods including the Kaplan-Meier approach. Proportional hazards modeling was used for multivariable analysis. RESULTS: Of 613 cohort members, 347 had NSEs, of whom 294 (85%) were followed ≥10 years (maximum = 17.9). A total of 170 in 294 (58%) achieved complete remission, 10 of whom (6%) relapsed. Seizure outcome at 2 years (remission, pharmacoresistant, unclear; p < 0.0001) and underlying cause (p < 0.0001) distinguished groups with complete remission ranging from ~20% to ~75%. Older age at onset was independently associated with a poorer chance of complete remission. Relapses occurred up to 7.5 years after attaining complete remission and were marginally associated with underlying cause (p = 0.06). INTERPRETATION: Complete remission occurs in over one-half of young people with NSE and generally persists. Meaningful but imperfect predication is possible based on underlying cause and early seizure control. The finding of age effects may play a role in meaningful identification of phenotypes, which could become fruitful targets for genetic and imaging investigations in these otherwise poorly differentiated epilepsies.
Commentary
The study by Berg et al. presents the risk and risk factors for complete remission in children without electroclinical syndromes, developmental encephalopathies, and potential electroclinical syndromes that could not be well classified who were observed for at least 10 years. Such nonsyndromic epilepsy represents almost 60% of childhood-onset epilepsy. Taking this interesting approach permits an evaluation of the cumulative risk for remission, risk factors, and risk for relapse absent children with electroclinical syndromes, in which the prognosis is part of the definition. The remaining cohort of nonsyndromic epilepsy is a mixture of epilepsy of unknown cause and epilepsy caused by structural or metabolic factors and more closely mirrors epilepsy in adults, permitting better comparisons of risk factors for remission across the lifespan. Restriction in this manner also hones the risk factors for remission that are identified, providing more valid estimates of the magnitude of risk associated with these factors.
In this study, remission is defined as at least 5 years free of seizures off medication and is called complete remission, not because it is the final endpoint for all children meeting criteria, but because it is a clinically meaningful endpoint that endures for most until the current time. This definition of remission has been used before by others, although it has not previously been called complete remission (1, 2). The more commonly used definition is 5 years free of seizures with or without medication (1–6), perhaps because, as pointed out in the paper, decisions to stop medication are multifactorial. In contrast, 5 years free of seizures can be observed uniformly across all people with epilepsy.
Another facet that separates this study from others is the availability of MRI exams in almost 90% of all nonsyndromic epilepsy observed for 10 years or more. Among these, 21% had an MRI abnormality. A prior publication of the Connecticut cohort has reported imaging in almost 80% of the full cohort, with abnormal imaging rare but present (3.5%) in children who were clinically normal and with abnormal imaging more common (12.7%) in the full cohort (7). These findings suggest that most, but not all, of the MRI abnormalities examined in the current report represent overlap with underlying brain disorder, abnormal neurologic examination, intellectual disability only, or autism spectrum disorder.
The occurrence of complete remission and risk factors for complete remission are examined in those with nonsyndromic epilepsy who were observed for 10 or more years. Others have also evaluated risk factors for remission in epilepsy cohorts observed for this long, including studies in Finland (5, 8), the United Kingdom (3), Holland (9), Canada (8), Sweden (4), and the United States (1); however, not all of these studies observed remission without medication or observed 5-year remission. Examining a minimum of 5 years in remission without medication instead of a minimum of 5 years regardless of remission on or off medication decreases the likelihood of relapse and thus improves the identification of risk factors for complete remission. This is true even if there is misclassification of children in 5-year remission on medication who could have achieved complete remission off medication but who were not withdrawn from medication because either they, their parents, or their physician did not support medication withdrawal. As previously reported by Annegers et al. (1), the risk for relapse in the Connecticut cohort was small—approximately 1% per year in those free of seizures for 5 years while off medication.
Early seizure outcome, evidence of underlying brain disorder, age at onset of epilepsy, status epilepticus, convulsive seizures, history of febrile seizures, first-degree family history of epilepsy, and initial seizure frequency were examined as risk factors for complete remission. In the final adjusted model, predictors of complete remission included evidence of underlying brain disorder, age <10 years at onset of epilepsy, and lack of early seizure remission. Not surprisingly, each of these risk factors has been identified in other studies of children (2, 8, 9) studies of children and adults (1, 3, 6). Importantly, other risk factors for 5-year remission on or off medication that were previously identified by others—such as seizure type (1, 3), epileptiform abnormality on EEG (2), lack of neonatal seizures (8)—were not identified in the unadjusted analysis of the current study. Their absence may be due to the longer length of follow-up of the current cohort, the focus on nonsyndromic epilepsy, or other unknown factors. The longer length of follow-up is an unlikely explanation that is supported by the comparisons made between those observed for at least 10 years and those not observed for as long.
The merit of this analysis is that it was designed to decrease the heterogeneity of examinations of remission in childhood-onset epilepsy, but heterogeneity remains because of the known strong associations between evidence of underlying brain disorder (i.e., structural or metabolic etiology) and decreased risk for remission (1, 10). This leads to an important question for future examinations of risk factors for remission. Do risk factors for complete remission (or remission generally) differ according to etiology of nonsyndromic epilepsy? This is explored briefly in the current paper during examination of the association between early seizure outcome and complete remission separately for those with underlying structural or metabolic etiology and those with unknown cause. Early seizure outcome is associated with remission when stratified by etiology, which is the same relationship observed in the combined etiologies. However, this may not be the case for all factors, such as status epilepticus, which was significant in unadjusted analysis but not in adjusted analysis, as well as for other risk factors examined that were not significant in unadjusted analysis that included all nonsyndromic epilepsy.
In the end, parents and children want to know if seizures will cease for a long enough time to be considered practically cured. The same care that is taken to create as meaningfully homogeneous a cohort as possible for the identification of genetic etiology should be taken for examination of epilepsy outcome. Studying cohorts of epilepsy more homogeneous for their etiology, and perhaps etiology and seizure type, may improve the identification of predictors of complete remission and remission generally. Numbers in any one study may be too small, but there remains potential for combining cohorts and determining seizure type and etiology by consensus to move further in the direction anticipated by this analysis. It is useful for us as researchers and may lead to potential interventions, such as more early surgery. Moreover, it is crucial for people with epilepsy and their families, who deserve to understand the course of the disorder.