The etiologic classification of epilepsy

Abstract

The classification of the epilepsies, and of epilepsy syndromes, has been a topic of much concern, and some controversy, for many decades. The 1989 International Classification of the Epilepsies (Commission on Classification and Terminology, 1989) was the latest attempt to categorize all the epilepsies according, essentially, to gross localization (partial vs. generalized) and gross etiology (idiopathic vs. symptomatic). These dichotomous concepts have been useful for teaching purposes, and for simplified discussions among physicians and even scientists, but we now know that epilepsy disorders cannot be clearly divided along these lines. With respect to location, all seizures begin somewhere, and even after they are fully developed, very few, if any, involve the entire brain. With respect to etiology, idiopathic has come to mean genetic, but inheritance patterns are complex and the mutations isolated for most of the idiopathic epilepsies are likely to represent susceptibility genes, rather than epilepsy genes. Susceptibility genes (and perhaps the same genes as for some of the idiopathic epilepsies) also contribute to the clinical manifestation of acquired epilepsies. No one knows what acquired factors may be necessary for the clinical manifestation of some genetic epilepsies. For these reasons, the International League Against Epilepsy (ILAE) Task Force on Classification (Engel, 1998; Blume et al., 2001; Engel, 2001, 2006), and later the Commission on Classification (Berg et al., 2010), charged over a decade ago with revising the 1989 classification, has maintained that any new classification should not be based on such false dichotomies. These concepts did have practical utility, however, and the Task Force and Commission were never able to devise a better way to categorize epilepsies based on location and etiology, nor were they able to find more accurate descriptors to replace the terms focal and generalized. As discussed by Berg & Scheffer (2011), however, genetic and structural/metabolic now replace idiopathic and symptomatic, but not as a rigid dichotomous concept (Berg et al., 2010). Another obstacle to creating a new comprehensive classification of the epilepsies has been the fact that although there is agreement on a number of epilepsy syndromes as unique diagnostic entities, most of which were formally classified as age-related and idiopathic, no one has yet come up with a simple way to divide what were considered to be focal symptomatic epilepsies into definitive syndromes that can be categorized in a clinically useful manner. Even mesial temporal lobe epilepsy with hippocampal sclerosis is no longer considered to be a single syndrome. Consequently, these conditions were not classified in the 2010 report (Berg et al., 2010). The 2010 report of the ILAE Classification Commission overcame the concern about false dichotomies by merely listing epilepsy syndromes by age of onset (Berg et al., 2010). By redefining the idiopathic epilepsies as genetic, they mean that the genetic mutation is directly responsible for the epilepsy and not responsible for some structural or metabolic disturbance, which then causes the epilepsy, but not always. By replacing the term symptomatic with structural/metabolic, they reinforce a desire to "tell it like it is." The final decision was essentially to limit the classification to generally accepted syndromes, which are almost all pediatric age-related genetic epilepsies. Although the problem of classifying structural/metabolic focal epilepsies was merely finessed by not including them in the proposed classification, the need to utilize a variety of factors, including etiology, to categorize these conditions into unique diagnostic entities is a work-in-progress, as noted by Berg & Scheffer (2011). The 2010 report has obviously generated some discontent, in part because clinicians, including epileptologists, have found it difficult to discard the simplistic dichotomies of the past, and also because most adults with epilepsy have yet to have their conditions represented in the new classifications. The thoughtful issues raised by Shorvon, and his proposed etiologic classification, deserve discussion at several levels (Shorvon, 2011). Is this a step forward or a step back? The first level concerns whether the concepts of idiopathic and symptomatic should have been abandoned, and whether we should not consider reviving them. The term idiopathic, which derives from the Greek word idio, meaning self, refers to epilepsy itself, sui generis, and not cause unknown, which is the way the term idiopathic is used in the rest of medicine. This has created some confusion in the past. Symptomatic is certainly a good descriptive term. Previous attempts at classification have used the terms primary and secondary, and what epileptologists call idiopathic is called essential in some other diseases. If we know exactly what we mean by idiopathic and symptomatic, I personally would have no strong objection to using them in place of what the latest classification report has suggested be called genetic and structural/metabolic; however, the concern is not the terminology, but the use of a dichotomous classification that implies that these terms are exclusionary. There is no justification at present to divide all known chronic epilepsy conditions into two classes. On the other hand, it is useful to continue to use the term idiopathic when describing, for instance, the idiopathic generalized epilepsies for historical reasons. Similarly, it is useful at times to refer to focal symptomatic epilepsies as a group when referring to conditions that are clearly due to structural/metabolic disturbances. However, referring to conditions like West and Lennox-Gastaut syndromes as symptomatic generalized epilepsies is problematic because some of these might be genetic. A case in point is Dravet syndrome, which, only a few years ago, would have been designated symptomatic. So at this level, perhaps we could retain the terms idiopathic and symptomatic for selective use. To return to a dichotomy, however, would be a step back. The next level concerns the need for a category of cryptogenic, which was in the 1989 International Classification of the Epilepsies (Commission on Classification and Terminology, 1989), but subsequently abandoned because there was disagreement about its definition. Cryptogenic was initially introduced because the term idiopathic was usurped and could no longer serve to define conditions in which the cause is unknown, as in the rest of medicine. Cryptogenic, which essentially means that the cause is hidden, was initially used to define conditions that are probably symptomatic, but for which no etiologic agent has been identified. Subsequently, the ILAE Commission on Epidemiology decreed that the term cryptogenic should be used when it is unknown whether a condition is idiopathic or symptomatic (Commission on Epidemiology and Prognosis, 1993). Consequently, the ILAE Task Force and Commission on Classification have discouraged the use of this term to avoid ambivalence, and again recommended to "tell it like it is" and merely say etiology unknown (Berg et al., 2010). Because Shorvon has defined cryptogenic as symptomatic with unknown etiology, this actually should be a subset of symptomatic in his classification. For a diagnostic manual there must be a category for individual patients for whom the etiology of their condition is unknown; however, an unknown category is unnecessary if we agree that not all conditions need to be dichotomized. Resurrecting cryptogenic would be a step back. The third level concerns the use of a category of provoked to refer not only to epilepsies, but more often to epileptic seizures that result from transient insults, where there is no enduring epileptogenic abnormality, and thus they would not be considered epilepsies (Fisher et al., 2005). The fact that epileptic seizures can be provoked under various circumstances does not reflect an etiology in the same sense that the concepts of idiopathic and symptomatic or genetic and structural/metabolic do. For the reflex epilepsies, conditions where seizures occur only with provocative stimuli, some, like reading epilepsy, or idiopathic photosensitive occipital lobe epilepsy, are genetic, whereas others, such as musicogenic epilepsy and startle epilepsy, are due to structural/metabolic causes. Consequently, there is no reason for categorizing provoked epilepsy separately, and there is no justification for including a classification of provoked seizures that are not epilepsy in a classification of epilepsy. With respect to some specific comments, I would disagree with Shorvon's comment that "In the genetic epilepsies, we are currently closer to defining cause by molecular mechanisms, but in the symptomatic epilepsies, the mechanisms of ictogenesis are largely obscure." Although we have made great progress in identifying molecular disturbances in both the genetic and the symptomatic epilepsies, which have become potential targets for antiseizure interventions, ictogenesis itself results from network aberrations and we have no idea in the genetic epilepsies (with the possible exception of typical absences) how these molecular disturbances result in epilepsy at the systems level, and only a rudimentary understanding has been elucidated for some of the symptomatic epilepsies, such as mesial temporal lobe epilepsy with hippocampal sclerosis. In any event, Jackson's view that most seizures have the same final common pathway is clearly incorrect. With respect to the specific classification that Shorvon has proposed, the idiopathic epilepsies essentially reproduce almost all of the syndromes included in the classification of the 2010 report (Berg et al., 2010), and dividing these into epilepsies due to a single gene disorder and epilepsies with complex inheritance is a reasonable contribution. With respect to symptomatic epilepsies

Keywords

Related Publications