The term temporal lobe epilepsy (TLE) has been used for many years to denote a variety of conditions associated with complex partial seizures of presumed temporal lobe origin. The 1985 International Classification of Epilepsies and Epileptic Syndromes²¹ divided symptomatic localization-related epilepsies according to the cerebral lobe of origin and recognized a syndrome of temporal lobe epilepsy defined purely on an anatomic basis. However, anatomic localization was usually determined purely from electroencephalographic (EEG) evidence, which could be misleading.³⁹

It has also been known for many years that the most common pathologic substrate of mesial TLE (MTLE), as loosely defined in the 1985 International Classification, is hippocampal sclerosis (HS).⁵ Data have accumulated in recent years strongly suggesting that MTLE with HS (MTLE-HS) represents a discrete syndrome that can usually be recognized in life.^{33,47,127,133,138} The International League Against Epilepsy (ILAE) recently convened a workshop of experts to discuss the natural history, pathologic features, pathogenesis, electroclinical, neurophysiologic, neuropsychological, structural, and functional imaging findings of MTLE-HS and to determine whether it should be considered a disease, a single syndrome, or a group of clearly distinct syndromes.¹³⁴ Their conclusion was that MTLE-HS should be considered a subtype of a greater syndrome of MTLE due to many causes, and that much more work is needed before MTLE-HS could be specifically characterized. Diagnosis of MTLE, and specifically MTLE-HS, early in the course of the disorder, however, is now considered important because disabling seizures and their consequences can be eliminated by anteromesial temporal lobectomy in 60% to 80% of patients, and early surgical intervention provides the greatest opportunity for complete psychosocial rehabilitation.^{34,127,130,133}

Bouchet and Cazauvieilh¹¹ first described the association between epilepsy and a sclerotic hippocampus in 1825, based on gross pathologic examination of brains from patients with “mental alienation seizures.” These lesions were believed to be an effect, rather than a cause, of epilepsy, which at that time referred only to generalized convulsions believed to originate in the medulla oblongata. In the latter part of the 19th century, however, Hughlings Jackson^63,64 recognized that partial seizures also represented epileptic phenomena and made an association between limbic-type seizures, which he called “intellectual auras” or “dreamy states,” and lesions in mesial temporal structures. During the same period of time, the neuropathologists Sommer¹¹⁷ and Bratz¹² suggested that HS might be an epileptogenic lesion.

With the advent of EEG, Gibbs et al.⁵² reported an ictal discharge pattern that they considered characteristic of psychomotor seizures; however, because they referenced all their electrodes to linked ears, they believed this phenomenon to be generalized. It was Jasper et al.^65,66 who pointed out that interictal and ictal epileptiform EEG abnormalities in psychomotor dreamy states and other phenomena now known to be limbic seizures originated in mesial temporal structures. Bailey and Gibbs⁶ were the first to perform anterior temporal lobectomy on the basis of EEG evidence alone, but it was the en bloc resection of Falconer,⁴³ which included mesial temporal lobe structures, that made it possible to perform systematic pathologic analysis of resected tissue. Not only was HS identified as being present in a high percentage of patients with medically refractory temporal lobe epilepsy, but also surgical outcome in patients in whom this pathologic abnormality was demonstrated was exceptionally good, supporting the argument that this structural lesion was the cause, and not the effect, of recurrent epileptic seizures. Falconer^44,45 also recognized the association between HS and both febrile convulsions and a family history of epilepsy, which suggested the existence of a specific syndrome.

In large part as a result of the success of EEG monitoring in identifying mesial temporal epileptogenic tissue, anterior temporal lobectomy became, and has remained, the most common surgical treatment for medically refractory epilepsy. Crandall et al.²⁴ subsequently took advantage of the en bloc resection and long-term depth-electrode recording to initiate basic research on temporal lobe epilepsy.³⁷ HS is now the most studied epileptogenic lesion, and thus the fundamental neuronal mechanisms responsible for MTLE-HS are the best understood of any form of human epilepsy. The elucidation of a unique pathophysiology coupled with characteristic clinical features and an increasing ability to identify hippocampal disturbances noninvasively with functional and structural imaging have all helped to establish MTLE-HS as a discrete epileptic syndrome.^22,134

Definitions of temporal lobe epilepsy and temporal lobe seizures have varied through the years, and confusion persists concerning their usage.³⁵ The terms temporal lobe seizure, psychomotor seizure, and limbic seizure are often used interchangeably to denote seizures with signs and symptoms that derive from ictal activation of mesial temporal limbic structures. According to some workers, these terms are reserved only for seizures that originate in mesial temporal structures; other investigators, however, interpret the terms more loosely and use them also to define seizures that might originate in temporal or extratemporal neocortex but rapidly and preferentially propagate to mesial temporal structures.³¹

Whereas the 1970 International Classification of Epileptic Seizures introduced the term complex partial seizure in
place of temporal lobe seizure,⁴⁹ the 1981 revision of this classification²⁰ redefined simple and complex partial seizures phenomenologically such that the latter term would be used to describe any partial seizure associated with impairment of consciousness. Because partial seizures can be associated with altered consciousness without involvement of mesial temporal limbic structures, not all complex partial seizures are temporal lobe seizures. Furthermore, because ictal discharges limited to one mesial temporal limbic area can produce characteristic symptoms such as epigastric rising, emotional and psychic experiences, and olfactory hallucinations during clear consciousness, temporal lobe seizures can also be simple partial seizures.⁵⁴ Consequently, whereas the 1981 International Classification of Epileptic Seizures has had definite clinical utility, the ILAE is now recommending that the concept of simple and complex partial seizures be abandoned. Seizures associated with MTLE are more correctly referred to as limbic seizures, whether or not consciousness is impaired.³⁶

The 1985 International Classification of Epilepsies and Epileptic Syndromes divided partial epilepsies according to the cerebral lobe of onset and recognized a number of subcategories of temporal lobe epilepsy based on characteristic patterns of ictal onset and propagation as recorded with depth electrodes.^21,131 Because of controversy regarding the precision of such electrophysiologic localization, the 1989 classification²² abolished anatomic categorizations but retained a syndrome of temporal lobe epilepsy based on associated clinical features, such as increased incidence of febrile convulsions and family history of epilepsy, as well as characteristic anterior temporal interictal EEG spikes and temporal lobe hypometabolism on positron emission tomography (PET). The term temporal lobe epilepsy, however, did not account for various etiologic factors and was used to denote the condition of patients both with and without apparent structural lesions. In some cases, patients with a diagnosis of temporal lobe epilepsy could conceivably have extratemporal epileptogenic regions that preferentially propagated to mesial temporal structures, producing characteristic temporal lobe seizures.

Before the advent of high-resolution magnetic resonance imaging (MRI), the term cryptogenic temporal lobe epilepsy had been frequently used to describe the condition of patients with the characteristic features of MTLE but no obvious lesions on structural imaging. Pathologic studies of mesial temporal tissue resected from patients with cryptogenic temporal lobe epilepsy revealed HS in most, so that the adjective cryptogenic, as opposed to lesional, was used tacitly to denote a subtype of MTLE that has now been more clearly defined as the syndrome of MTLE-HS.

Ammon’s horn sclerosis and mesial temporal sclerosis are the two most common pathologic terms that have been used more or less synonymously with HS, although strictly speaking they imply different degrees of anatomic involvement.⁴ The term hippocampal sclerosis refers to a specific type of hippocampal cell loss involving CA1 and hilar neurons most and CA2 neurons least, which distinguishes this entity from nonspecific cell loss with other causes.⁵ Other characteristic features, such as mossy fiber sprouting¹²³ and selective loss of somatostatin and neuropeptide Y–containing hilar neurons,²⁵ also help to identify this distinct pathologic entity.

Because the syndrome of MTLE-HS has only recently been clearly defined, and because only the medically refractory forms of this disorder referred to epilepsy surgery centers are usually identified, no epidemiologic information is available. It has been reported that 40% of patients with epilepsy experience complex partial seizures⁵⁰; however, not all these patients necessarily have temporal lobe epilepsy, let alone MTLE-HS. In surgical series, careful pathologic analysis of hippocampal specimens, including cell counts and special staining procedures, reveals that as many as 70% of patients with medically refractory temporal lobe epilepsy may have HS,⁵ indicating that the vast majority of patients with medically refractory temporal lobe epilepsy have MTLE-HS. Because of the probable increased incidence of a family history of epilepsy among patients who have MTLE-HS,¹³³ there may be some relationship between this syndrome and the recently described benign familial form of MTLE, which can also be associated with HS.^10,16 Although it is currently impossible to estimate the number of patients with MTLE-HS whose seizures are adequately controlled by antiepileptic drugs, two studies provide some evidence from nonsurgical series evaluated with high-resolution MRI. Semah et al.¹¹⁴ reported on patients seen in a Paris epilepsy clinic over a 7-year period, approximately one half of whom had a diagnosis of temporal lobe epilepsy. Of these, half had evidence of hippocampal atrophy on MRI. At least one quarter of their population, therefore, could be given a diagnosis of MTLE-HS, and the percentage was probably higher because not all patients with this condition have sufficient atrophy to be visible on MRI. More importantly, of all the diagnostic categories, the most refractory to antiepileptic drugs was MTLE-HS. Only 11% of patients with this diagnosis, and only 3% with dual pathology (that is, HS and some other MRI lesion), had been seizure free in the previous year. This report suggests that MTLE-HS may be the most common, and most medically refractory, form of epilepsy. Another report from Glasgow by Stephen et al.,¹²¹ however, found a similar proportion of patients with MRI evidence of HS, but only 46% of these were medically refractory (although this group was still the most intractable). The difference may be due to the fact that the Paris clinic included tertiary referral patients. The time at which patients with MTLE were referred may also influence the prevalence of intractability, because this condition typically has a stuttering course with long periods of remission. Berg et al.⁸ reported on a large cohort of patients undergoing surgical treatment for medically refractory focal epilepsy, most of whom had MTLE, and found that it took an average of 9 years for the establishment of medical intractability.

The epileptogenicity of this disorder results from loss of specific neurons in the hippocampus and synaptic reorganization of surviving cellular elements to cause hypersynchronization and hyperexcitability. The characteristic features of cell loss and reorganization, as well as the electrophysiologic aberrations that ultimately give rise to spontaneous seizures, are described in detail in Chapters 13 and 41. Although the pathophysiology of HS in fully developed MTLE-HS has been relatively well defined by studies of patients in epilepsy surgery centers and some very good experimental animal models exist, the events that initiate this process in humans are as yet unknown.

An association between mesial temporal sclerosis and febrile seizures was reported more than 30 years ago,⁴⁴ suggesting a causal relationship. Subsequent studies in both animals and humans have challenged this concept; some patients with MTLE-HS do not have a history of febrile convulsions,⁴⁷ and prolonged seizures in immature animals do not produce HS.⁹⁴ There are methodologic concerns with these studies, however, and more recent series of patients having MTLE-HS report a strong association with early risk factors, with as many as 66% of patients having prolonged febrile convulsions^{14,15,47,60,61,115,116,138} and a variety of other early insults, such as trauma and infection.⁸³ Furthermore, there is
evidence that the process, once begun, does not stop after seizures appear but rather continues to progress over time.^47,82

Also of etiologic importance is the observation that HS is sometimes associated with microdysgenesis⁸⁵ and occurs in patients with dysplastic lesions, such as hamartomas and heterotopias, but only rarely in patients with neoplasms.⁷⁹ This association, coupled with the increased incidence of a family history of epilepsy and febrile convulsions, suggests a genetic or congenital predisposition to epilepsy. Consequently, familial or congenital disturbances could be a necessary precursor that permits early noxious insult to give rise to the characteristic cell loss and neuronal reorganization necessary for the development of MTLE.

The current characteristic clinical features of MTLE-HS are derived almost exclusively from patients with medically intractable seizures who are being evaluated for surgical intervention. Onset of habitual seizures typically begins toward the end of the first decade of life, and seizures often initially respond well to antiepileptic drug treatment. Characteristically, patients do well for several years, but seizures then return in adolescence or even early adulthood and become refractory to medical treatment. Of these patients with medically intractable seizures, the great majority have a history of complicated febrile convulsions or other initial precipitating injury.^47,83