More than 2000 years ago, epilepsy was recognized as a physical disease arising in the brain. Despite the enlightened approach reflected by the Hippocratic authors, the prevailing views about epilepsy were that it was due to supernatural forces—demonic possession or divine retribution. Some held that epilepsy was contagious. These attitudes largely account for the tremendous stigma associated with epilepsy, a stigma that still lingers to a greater or lesser extent.

The establishment in the mid- to late 1800s of specialized centers for the care and treatment of “epileptics” represented a shift from the highly stigmatized views of epilepsy as due to evil, supernatural forces to one in which epilepsy was regarded as a medical disorder. With the concentration of people with epilepsy in these specialized centers, early neurologists began to describe epilepsy from the medical perspective. This led to the characterization and long-held belief that epilepsy was a uniformly progressive, nonremitting disorder. Sir William Gowers wrote one of the most famous and influential dissertations on epilepsy, in which he said, “The tendency of the disease is toward self-perpetuation; each attack facilitates the occurrence of another by increasing the instability of the nerve elements…. The spontaneous cessation of the disease is an event too rare to be reasonably anticipated.”²⁸ Since then, our understanding of how to study the natural history of a disorder and of epilepsy in particular has developed. This has been in large part due to conceptual and methodological contributions from two very different areas—epidemiology and “syndromology.”

The first development to be described here was greatly facilitated by the advent of modern epidemiologic and statistical principles and techniques and their application to the study of epilepsy in the overall population. This helped to lead to the recognition that findings based on the most severely affected patients seen in tertiary referral centers were of limited value in treating and counseling the vast majority of patients in the population, whose epilepsy was often mild enough not to require treatment at such centers. In other words, the earlier observations suffered from multiple and severe forms of bias.⁴⁹ The early epidemiologic efforts were largely due to Mayo Clinic investigators led by Leonard Kurland and later by W. Allen Hauser and J. Fred Annegers.^30,31,39 The epidemiologic approach cast a broad net and provided invaluable information about the frequency, risk factors, and overall outcomes in the population at large.^1,2,3,6,7

The early epidemiologic studies yielded valuable information about the prognosis of epilepsy. They provided an overview by providing an understanding at the population level of the frequency of epilepsy, the distribution by age at onset and gender, and an assessment of the various factors that cause it. In general, however, epilepsy itself was treated as relatively undifferentiated disorder. Although the epidemiologic studies identified some broad categories with important prognostic significance (e.g., presence of an underlying symptomatic factor likely responsible for the epilepsy, the age at onset, and whether the seizures were of generalized or focal onset), these were viewed as factors that modified a single disorder much as height and weight can vary among members of a single species.

The other major development that occurred was the recognition that epilepsy, much like cancer, was not a single disorder but a term used for a wide range of disorders. The endeavors in this area have been and still are at the heart of efforts to identify and investigate individual forms of epilepsy, often called syndromes. This movement has largely come from the European schools of pediatric neurology. Among these the French investigators have been the most prolific and productive, with many notable and important contributions coming from German, Italian, and Japanese investigators as well. Efforts to apply this approach to adult-onset epilepsy are hindered perhaps by the relative lack of distinctive diversity within these forms of epilepsy.

The concept of an epilepsy syndrome was defined in 1985 as “an epileptic disorder characterized by a cluster of signs or symptoms customarily occurring together. The signs and symptoms may be clinical or findings detected by ancillary studies. In contradistinction to a disease, a syndrome does not necessarily have a common etiology or prognosis. Some epileptic syndromes are, however, of great prognostic importance.”¹⁸ An updated definition was provided recently: “A complex of signs and symptoms that define a unique epileptic condition.”²⁵

Contrary to the epidemiologic approach, individual forms of epilepsy, from this vantage point, represent complete, integrated disorders. Epilepsy is not a single disorder with a variation in prognosis statistically defined based on circumstances and risk. Traditionally, the epidemiologic approach would require an individual to have had at least two unprovoked seizures on separate days for the condition to qualify for the label of epilepsy.²⁰ From the syndromic perspective, it is not clear why it is necessary to withhold the diagnosis of the epileptic disorder in someone who, for example, has had a
single unprovoked seizure and whose underlying disorder can be clearly identified or in a child who presents with febrile status epilepticus and is shown to have a functional SCN1A gene mutation. In fact, to the extent that early identification of the nature of the disorder influences choice of treatment and management options, and to the extent that the correct choice may lead to better outcomes, it seems it would be irresponsible to insist on waiting for two unprovoked seizures to diagnosis and treat the underlying epileptic disorder.

With a syndromic approach, a disorder such as febrile seizures, which, for good practical reasons, has been distinguished from epilepsy proper,^{6,21,41,42,43} might reasonably be reintegrated into the spectrum of epilepsy. The same might be said of benign nonfamilial neonatal seizures. In doing this, however, the older notion of epilepsy as a uniformly nonremitting, progressive disorder must be abandoned and the very broad spectrum of manifestations, duration, outcomes, and consequences of the wide variety of epilepsies acknowledged. In particular, without denying or forgetting the severe end of the epilepsy spectrum, we need a widespread recognition that, at least from the perspective of seizure control, many forms of epilepsy are actually quite mild, easily treated, often resolve on their own, and are associated with minimal consequences.

In the 1989 proposal for the classification of the epilepsies, two key axes were identified, one for generalized versus localization-related conditions and another for etiology (idiopathic, symptomatic, and cryptogenic).¹⁹ Currently, these are well-accepted terms for organizing syndromes and epilepsies.