Classification concepts and terminology: Is clinical description assertive and laboratory testing objective?


  • The following commentaries were invited by the Editors-in-Chief

Our understanding of epilepsy has changed remarkably over the last 20 years. Modern imaging, genetics, molecular biology, and cellular physiology have introduced considerable new information and, to some extent, modified concepts. This issue of Epilepsia publishes a report of the ILAE Commission on Classification and Terminology containing a number of suggestions for revised terminology and concepts. The authors’ intention—to “bring epilepsy out of the shadows of expert opinion and assertion-dominated arguments”—is certainly laudable. One wonders, however, if previous efforts of classification had failed to achieve, or at least address, a similar objective. Had classification been relegated to such an unacceptable position by previous inaccurate work? Was scientifically supported knowledge at the time of previous classification proposals insufficient to do better? Is a clinical classification destined to rely on speculative inference? There are no firm or satisfying responses to these questions. Realizing an ideal, scientifically sound, and unassailable clinical classification is probably a utopia. Such a classification might be possible for some patient subgroups, where molecular genetics, imaging, or neuropathology can validate (always in retrospect) clinical hypotheses—but it is by no means realistic in the clinical environment at large. An effort to impose order onto a domain as complex as epilepsy is naturally conditioned by different approaches and priorities, even more so given an international perspective. Compromises are necessary, as should be acknowledged as an unavoidable limitation to any future classification project.

Conceivably, a first “scientific” step toward building a classification system would be identifying its main use. One would expect that the key purpose for a Classification produced by the International League Against Epilepsy (ILAE) would be to improve care for the highest number of patients with epilepsy around the world. Within this context, what is the minimum information necessary to classify most patients within meaningful categories, and what is the maximum possible degree of accuracy to classify them accurately? This Commission report seems to have addressed more the second question than the first, without considering that one may contain the other, if the classification scheme is properly built. A classification that focuses on internationally distributed realities, and is organized according to a module where increasing information is exploited to progressively improve patient diagnoses within more specific categories, would be highly functional and accessible to a wide public. For example, in countries with limited imaging resources [e.g., where brain magnetic resonance imaging (MRI) is not available], or for those patients for whom the nature of the lesion is uncertain even after high-resolution imaging, the descriptor “symptomatic temporal lobe epilepsy” might represent the basic conceptual class; the more accurate descriptor “epilepsy with focal seizures secondary to cortical dysplasia in the temporal lobe” would constitute a further step.

Most concepts that formed the body of the 1989 Classification proposal (Commission on Classification and Terminology of the International League Against Epilepsy, 1989) were derived from the descriptive attitudes of European epileptologists, and initially encountered some resistance upon entering the English-speaking world. However, a series of important epidemiologic, scientifically sound studies conducted in the United States, the so called “Connecticut study,” provided the 1989 proposal with a solid, retroactive scientific background. In a prospective, community-based study, Berg et al. (1999a), using the 1989 Classification criteria and based on information available at diagnosis or generated as part of the initial assessment, found a high degree of interrater agreement In a companion paper, Berg et al. (1999b) prospectively studied 613 children with newly diagnosed epilepsy and estimated that at the time of diagnosis, epilepsy syndromes were classifiable in all but four children. In a subsequent follow-up study, the same group (Shinnar et al., 1999) prospectively recruited 407 children with a first unprovoked seizure and followed them for a mean of 9.4 years. In 182 children with two or more seizures, etiology and epilepsy syndromes were classified using the 1989 ILAE classification. “Factors associated with a favourable prognosis included an idiopathic or cryptogenic etiology and having a localization-based idiopathic epilepsy syndrome.” The authors concluded that “after two seizures, childhood-onset epilepsy can be classified by etiology and epilepsy syndrome.” In a fourth study, Berg et al. (2000) compared the classification of epilepsy syndromes in a cohort of children with newly diagnosed epilepsy, as made on the basis of information available at diagnosis versus that made 2 years later, and concluded that “epilepsy syndromes can, for the most part, be identified at the time of initial diagnosis.” Two years later, rectifications were made in only 9.8% of cases, and most of these involved syndromes that represented incomplete classifications in the first place. The ILAE Classification of the epilepsies was, therefore, judged as a meaningful tool for epidemiologic studies of newly diagnosed pediatric epilepsy.

All these studies have produced firm evidence that the main syndrome categories introduced in the 1989 Classification proposal are reliable for clinical and epidemiologic purposes. However, improvements have been introduced as uncertainties emerged about some terms and categories. We now understand, for example, that the term “cryptogenic” conveys ambiguous implications and should not be used (Engel, 2001). One might also question how specific each syndrome category is (for example, absence epilepsy as a continuum vs. age related absence epilepsy syndromes), and to what extent prognosis is conditioned by etiology rather than by the syndrome itself (West syndrome for example) (Guerrini, 2006). New findings, particularly in the genetic domain, have had important implications. In particular, the characterization of monogenic epilepsies has opened new perspectives for promoting further research; to date, however, these genetic insights have had a direct impact on only a limited number of patients—not sufficient justification to discard the main conceptual groupings. The categories of idiopathic and symptomatic still have solid implications for diagnosis, management, and prognosis, and these categories allow a large number of patients to be rapidly framed within meaningful groupings.

The three part categorization—Genetic – Structural/Metabolic – Unknown—introduced in this Commission report is of great interest and represents an innovative approach. However, some recently characterized genetic disorders challenge the distinction between the “pure” genetic epilepsies, which are ideally epitomized by ion channel alterations, and the Structural/Metabolic disorders in which a separate abnormality is interposed between the genetic defect and the epilepsy. For example, epilepsy in nonmalformation ARX phenotypes (at least in relation to some of the mutations of this gene) or in patients with STXBP1 or CDKL5 mutations/deletions, is difficult to assign to either the “genetic” or the “structural” category. It seems to be a primary expression of the genetic defect: since all patients have severe seizures but no structural lesion is recognizable as far as the diagnostic dimension can be pushed in vivo. Neuropathology and molecular pathology will hopefully provide an answer to these queries in the future.

When looking at this Commission report in its entirety, it appears that suggestions to change terminology and concepts (other than nomenclature) or changes to the list of epilepsy syndromes are rather limited. A number of concept categories that had already been recognized and updated in the 2001 and 2006 Task force reports (Engel, 2001, 2006) have been reinforced over time, after attention was drawn to them, and are a solid reference point for the clinician. As stated in the Commission report, current knowledge is not yet adequate to propose a new Classification of the epilepsies.


I confirm that I have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. I have no conflicts of interest to disclose.