Epigraph Vol. 21 Issue 2, Spring 2019

The ILAE seizure and epilepsy classifications: Critique, response and the way forward

Linnaeus is renowned for devising a biological taxonomy in which to classify all living things. And while the framework is clear, the act of classifying organisms within that framework is not. The more detailed the classification becomes, the more complex and confusing it can get.

The idea of a “species” is something that everyone learns but no one can clearly define. Even Charles Darwin, who penned “On the Origin of Species”, had a definition of “species” that was something like Supreme Court Justice Potter Stewart’s definition of obscenity: I’ll know it when I see it.

Many species are categorized by how they look; others, by their geographical range; still others, by how they mate. But the lines between species are fuzzy. The grade-school mantra that “two species cannot interbreed” has been disproven time and time again—certainly by plants and yeast, but also by animals. Animal hybrids have happened for millions of years—from the human/Neandertal pairings that led to modern-day Homo sapiens, to a warbler discovered last year in the United States as the offspring of parents from two different genera (Mom was Vermivora leucobronchialis, a Brewster’s warbler; Dad was Setophaga pensylvanica, a chestnut-sided warbler).

The advent of rapid DNA sequencing adds even more complexity to the issue. If differences in DNA define species, how many differences should there be, and what kinds of differences matter?

Regardless of the issues and problems with biological taxonomy, it remains a common language around the world to help classify and discuss living things. Its users work within it, adjusting as necessary.

Classifications and critiques

Robert S. Fisher
Robert Fisher

By a similar token, the ILAE operational classifications of seizures and the epilepsies aim to provide a common language worldwide, to aid clinicians in diagnosing and managing patients with epilepsy, to allow clinicians and families to communicate more clearly, and to be of value for research and scientific study. Though not taxonomies, per se, the ILAE classifications provide frameworks suitable for use by a wide range of medical professionals, without the need for expensive technology.

After years of work—including in-person meetings, hundreds of emails and multiple drafts—ILAE’s most recent Operational classification of seizure types, and its classification of the epilepsies, were published in 2017. Earlier this year, both classifications were critiqued in an article by Hans Lüders and colleagues, published in Epilepsia. The paper touched on several areas of criticism, which were addressed in a recent response by Fisher and colleagues (also published in Epilepsia). Lüders and colleagues also published a proposed four-axis classification system in early 2019.

The general theme of the criticism was that the ILAE classifications lacked detail and the framework was less than ideal. Lüders also noted a lack of testing. “You should develop a system and test it for a period of time, and then release it,” he said. “[The ILAE classifications were] like writing a computer program and just releasing it without testing it.”

Differing perspectives

Ingrid Scheffer
Ingrid Scheffer

“I think his point of view is very specialized,” said Ingrid Scheffer, who led the effort on the classification of the epilepsies. “He is a leader in epilepsy surgery evaluation; he’s surgically focused, focused on finding the epileptogenic zone. His concepts are great for epilepsy surgery patients, but not appropriate for the common person dealing with epilepsy in an under-resourced place in Africa or a general practice in Australia.”

“A classification is a tool,” said Robert S. Fisher, who led the seizure classification task force. “We were trying to get a fairly globally descriptive system that could be used by patients and their families, clinicians, journalists, pharmaceutical companies and regulators – we wanted it to be broadly applicable.”

The epilepsy classification followed a similar goal, said Scheffer. “The epilepsy classification is important for every patient you see,” she said. “I use it every day with every patient. We wanted to use terms that parents, patients, general practitioners, pediatricians, neurologists could all use and understand

Criticism and response

Hans Lüders
Hans Lüders

Criticism: In both classifications, the definitions mix semiological terms with epileptogenic zone terminology. Lüders et al. felt this was confusing and suggested a pure semiological seizure classification as part of a multidimensional epilepsy classification, in which seizure semiology and epileptogenic zone are independent dimensions.

Response: Mixing semiological and electro-anatomical data was intentional. The ILAE Task Force thought it best to use all available information to classify a seizure. They considered a pure anatomical classification, as limbic, thalamocortical, cortical, brainstem and other seizures, but decided the scientific basis was still lacking.

The division into focal and generalized onset seizures has great practical value in directing the search for lesions, medications and surgical opportunities. Most people with epilepsy will never undergo video-EEG monitoring and will not require categorization of seizure type beyond focal versus generalized. The new “unclassified” category is a placeholder for a lack of information, rather than a true seizure type.

Criticism: Simple and widely accepted terminology has been replaced by complex terminology containing less information. “For example, the expression ‘aura’ is replaced by ‘focal aware seizure’,” noted the critique. “This change does not bring additional biological value but rather encumbers description.”

Response: “One of the things we set out to do was to use transparent language,” said Scheffer. “We wanted language that patients could understand, not just doctors.”

The term “aura” is problematic because most people, and some physicians, do not consider an aura to be a seizure. However, auras are seizures; the updated classification makes that clear.

The aim, said Fisher, was to create classifications that could be used by physicians and patients alike: a common language. Research shows that about half of patients do not use the same terms to classify their seizures as their physicians do, said Fisher. “That’s not a great situation. So we tried to improve those features.”

The 1981 classification of seizures—which was the basis for the 2017 update—used terms like “simple partial” and “complex partial”, which the Task Force changed to focal, focal aware, focal impaired awareness and cognitive.

We thought some of the terms [in the 1981 classification] were fairly obscure,” said Fisher. “Like ‘complex partial’ – some patients weren’t clear what that meant. And the term ‘simple partial’ was actually offensive to some of them, because there’s nothing simple about a seizure.”

Criticism: Seizure evolution cannot be described in any detail. “This … makes it impossible to express in detail the evolution of most focal epileptic seizures, and this important information is lost,” wrote Lüders et al. “A true semiological seizure classification allows neurologists to already have an anatomofunctional perspective of seizure onset and evolution when taking the clinical history in the clinic. Over time, this is likely to provide a gestalt for surgical candidacy from the moment refractoriness is declared.”

Response: Describing the evolution of seizures provides more information on the symptomatic zone, but not necessarily the epileptogenic zone. “We had to decide: Do you classify a seizure based on where it starts, or based on the most prominent feature once the seizure gets going?” said Fisher. “We decided on the former, because there’s something wrong with that part of the brain.”

Fisher et al. note that no classification system could classify every seizure evolution pattern. “With thousands of possible propagation patterns for seizures and corresponding sequences of behavior,” they write, “a new strategy will be needed to succinctly describe them all.”

Criticism: In the epilepsy classification, level two (epilepsy category) overlaps almost 100% with diagnostic level one (seizure type). Luders et al. propose a multidimensional epilepsy classification system, suggesting that one option be a four-dimensional system that includes seizure type (defined exclusively by seizure semiology), location of the epileptogenic zone, etiology, and comorbidities.

Response: Fisher noted that the Lüders suggestion of a multidimensional system is similar to the ILAE classification of the epilepsies, which includes several dimensions representing seizure type, epilepsy type (which takes into account multiple seizure types and evolution over time), etiology, comorbidities and syndromes. These are not hierarchical, but add complementary information important for management of the patient’s epilepsy.

Criticism: The design of different classifications with distinct frameworks for newborns, adults, and patients in status epilepticus is confusing. Luders et al stated that the classification of the epilepsies of Scheffer et al. calls for classification of progressively more detailed diagnostic categories (seizure type, epilepsy category, epilepsy syndrome, and epilepsy with specific etiology), while the status epilepticus classification of Trinka et al. is a multidimensional classification including: semiology, etiology, EEG, and age. They suggesting using the same framework for both adults and children, and also for patients with status epilepticus.

Response: The ILAE seizure classification Task Force developed a system applicable to children, adolescents, adults and seniors. There is ongoing work by an ILAE neonatal seizure commission; that classification will likely be largely consistent with the ILAE 2017 seizure classification.

Status epilepticus is something of a different animal, noted Scheffer. “You may have a seizure type within status, but the classification of status epilepticus has to do with the different stages and courses, not how it presents,” she said. “That is what affects how it is managed, so that’s why it was dealt with separately.”

Criticism: Epilepsy syndromes are less important than in the past. “All syndromes are by definition empirical and artificial,” wrote Lüders et al. “Modern diagnostic techniques including MRI and genetic testing now allow precise diagnosis of epilepsy causes, so identification of syndromes is less important than it once was, although several still impact therapy decisions (such as West syndrome, benign rolandic epilepsy, Dravet syndrome, juvenile myoclonic epilepsy) or have relevance to genetic research (such as Dravet syndrome).

The Lüders critique states that as diagnostic technology and knowledge about epilepsy improve, it is likely that more syndromes will become obsolete; therefore, a classification scheme should not preserve them. Instead, it should “define as precisely and objectively as possible the characteristics of each individual case of epilepsy to facilitate discovery of new etiologies.”

Response: The ILAE 2017 Epilepsy Classification also includes epilepsy syndromes, which are critical for understanding a patient’s epilepsy, especially for epilepsies beginning in childhood.

“Epilepsy syndromes are different from types,” said Scheffer. “Syndromology is useful if you look at epilepsy as a complex puzzle. A syndrome informs treatment, diagnosis and prognosis. Syndromes are mostly issues for pediatric epileptologists, since 80% of epilepsy begins in childhood or adolescence.”

Criticism: Everyone must use the ILAE classifications; Epilepsia will not publish articles that do not use them.

Response: Fisher et al. wrote: “The authors of the seizure classification paper did not invoke such a rule. The League's journals reasonably require that submitted papers use a common language when referring to seizures and epilepsy, so that information in published papers can be directly compared. The journals do not forbid concurrent use of other terms as well, so long as they are mapped (as much as is possible) to the ILAE classification. Without a common language, meta-analyses and other valid scientific reviews cannot reliably be performed.”

Change takes time

Both Fisher and Scheffer have the perspective that anything new will take time to work itself into common usage. “[The updated classification is] not perfect, but it’s being received pretty favorably,” said Fisher. For example, a study of 606 Norwegian children found the new classification provided a higher precision of epilepsy diagnosis, with a broad syndrome characterization in 93% and a defined syndrome in 37%. A retrospective review of 200 epilepsy patients found that the 1981 classification resulted in 89 unclassified seizure types; 75% of them could be classified using the updated seizure classification.

“You wonder when you do something like this if it’s going to work out, but it’s working out,” Scheffer agreed. “It’s a tool for us that takes current understanding into account.”

This presentation of different classification systems makes it clear that there are different ways to organize our thinking about seizures, depending on the purpose.  The physician treating the patient, the epidemiologist working to develop an overview of epilepsy in a population or a neurosurgery center working to have the best possible outcome from a surgical intervention may each want a classification tailored to their specific needs. 

“[Classification is] always going to be a moving target, and that’s okay,” said Scheffer. “The Holy Grail we’ve been wanting is a scientifically based classification. We’ll have that when we understand the brain—and we’re so far from understanding the magical mysteries of the human brain.”

In clinical practice, where the information is often limited and simplicity is essential, complexity is rarely possible. No doubt the systems will change again, as information is gathered about the utility and benefits of different methods, and more is learned about epilepsy. As the old saying goes: For everything there is a season — and for every classification there is a reason.

Changes in seizure type classification from 1981 - 2017