- Email: [email protected]
Temporal lobe epilepsy in infants and children
Brain & Development 20 (1998) 135–141
Review article
Temporal lobe epilepsy in infants and children Blaise F.D. Bourgeois* Washington University, Department of Neurology, Suite 12E25, St. Louis Children’s Hospital, One Children’s Place, St. Louis, MO 63110, USA Received 24 October 1997; revised version received 6 January 1998; accepted 20 January 1998
Abstract Temporal lobe epilepsy in adults and adolescents is a fairly homogeneous syndrome, both in terms of seizure semiology and in terms of its pathology, and it has been studied extensively. Temporal lobe epilepsy in infants and young children has begun to receive increasing attention in recent years, and a different clinico-pathological picture has emerged. Clinically, the concept of complex partial seizures, which may be useful in adults, is difficult to apply to infants, since it is often not possible to assess impairment of consciousness in this age group. The main distinctive features of complex partial seizures of temporal lobe origin in infants are (1) a predominance of behavioral arrest with possible impairment of consciousness, (2) no identifiable aura, (3) automatisms that are discrete and mostly orofacial, (4) more prominent convulsive activity, and (5) a longer duration (more than 1 min). In addition, seizures of temporal lobe origin in infants may appear clinically generalized, such as infantile spasms or generalized tonic seizures, or can occasionally represent a benign syndrome. The neuropathological findings of temporal lobe epilepsy in infants differ even more than the clinical seizure semiology. In contrast to adult and adolescent patients, mesial temporal sclerosis is a rare finding in infants, in whom the pathological abnormalities associated with seizures of temporal lobe origin consist mostly of dysplasias, migrational disorders, hamartomas, and low-grade tumors such as gangliogliomas. Mesial temporal sclerosis is seen more often in older children than in infants, and its pathogenesis remains a subject of controversy. 1998 Elsevier Science B.V. Keywords: Complex partial seizures; Temporal lobe; Infants; Children
1. Introduction Overall, among the epilepsies of adult and adolescent patients, the temporal lobe plays an extremely important role. In this age group, the temporal lobe is by far the most commonly involved lobe in partial epilepsies, and temporal lobe epilepsy represents a fairly homogeneous syndrome [1]. In the past decade, temporal lobe epilepsy and complex partial seizures in infants and children have been the focus of many clinical investigations and observations. From this work, a picture has emerged that suggests that temporal lobe epilepsy in infants and children represents a different clinico-pathological entity. The present review will focus first on the classical mesial temporal lobe epilepsy of the adult, then the concept of complex partial seizures in infants and children will be discussed
* Fax: +1 314 454 2523; e-mail: [email protected] 0387-7604/98/$19.00 1998 Elsevier Science B.V. All rights reserved PII S0387-7604 (98 )0 0010-2
and, finally, our current understanding of the clinico-pathological aspects of temporal lobe epilepsy in infants and children will be presented.
2. Mesial temporal lobe epilepsy syndrome Among adolescents and adults with the mesial temporal lobe epilepsy syndrome [1], there is an increased incidence of a family history of epilepsy and febrile seizures. A personal history of febrile seizures, which are often prolonged, has been reported in 9–50% of the patients. The onset of the nonfebrile seizures occurs mostly between the ages of 5 and 10 years. They are almost invariably complex partial seizures. Auras, most commonly with epigastric onset and a duration of 5 to 60 s, are experienced by about three-quarters of the patients. The actual complex partial seizure consists of an arrest of activity with a motionless stare, clouding of consciousness, and oroalimentary and hand automatisms.
136
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
There is at times dystonic posturing of the contralateral hand. The duration of the seizure is mostly between 30 and 90 s, with a postictal reorientation that may last several minutes. When spikes are captured on interictal scalp EEG recordings [2], they are usually anterior temporal, with maximum at the sphenoidal electrodes, ear electrodes, or anterior temporal electrodes (T1/T2, FT9/FT10). Contralateral temporal spikes can be seen in 50% of the patients. Ictal surface EEG seizure patterns usually consist of rhythmical activity in the delta, theta, or alpha frequency range, mostly in the ipsilateral temporal region or, less often, over the ipsilateral hemisphere. Invasive ictal EEG recordings reveal 13–25 Hz discharges in medial temporal lobe structures, amygdala and anterior hippocampus [3]. This ictal activity propagates initially to the ipsilateral temporal neocortex in 60%, and to the contralateral hippocampus in 30% of the cases [3]. The characteristic MRI findings of mesial temporal lobe epilepsy represent a triad which consists of (1) hippocampal atrophy on T1-weighted images, (2) an increased hippocampal signal on T2-weighted images, and (3) a loss of internal structure of the hippocampus (decreased white/gray differentiation) [4]. Interictal positron emission tomography (PET) scanning reveals a zone of ipsilateral temporal and extratemporal hypometabolism that is much larger than the EEG zone of interictal activity and ictal onset, and also larger than the zone of neuropathological changes [5]. The zone of hypometabolism involves the temporal lobe in a rather diffuse fashion and often also the ipsilateral hypothalamus. In a more recent study [6], depression of cerebral glucose metabolism was also found in the fronto-orbital cortex bilaterally, as well as ipsilaterally in the posterior insula and in the thalamus. In addition, patients with left temporal lobe epilepsy were found to have hypometabolism in the left inferior frontal gyrus and in the left superior temporal gyrus. No corresponding areas of hypometabolism were seen in patients with right temporal lobe epilepsy. No correlation could be demonstrated between the pattern of hypometabolism on PET scans and the neuronal density of resected hippocampi [7]. Single photon emission computed tomography (SPECT) scanning shows temporal hypoperfusion interictally, but interictal SPECT has a lower resolution than interictal PET scanning. Ictal SPECT scanning reveals temporal hyperperfusion during the seizure; immediately after the seizure, there is mesial temporal hyperperfusion and lateral temporal hypoperfusion [8]. The anatomy and pathophysiology of mesial temporal lobe epilepsy have been studied extensively [9]. Although recordings with invasive electrodes in humans suggest that seizures begin in the hippocampus, the clinical methods are relatively insensitive. It is conceivable that activity generated outside the hippocampus, such as in the pyriform cortex, may become detectable only after propagation and hippocampal amplification [10]. Anatomically, mesial sclerosis is characterized by hippocampal neuronal loss of
at least 30%, and the CA1 region is consistently reported as being the most affected [11]. It is possible that neuronal loss precedes the onset of clinical seizures, but ongoing seizure activity may contribute to the sclerosis. In specimens from 32 temporal lobectomies, there was a highly significant association between severe neuronal loss and early childhood seizures that were prolonged and/or lateralized [12].
3. Complex partial seizures in infants and children Since complex partial seizures are the predominant seizure type in the mesial temporal lobe epilepsy syndrome, it is appropriate to begin a discussion of temporal lobe epilepsy in infants and children with a review of the electroclinical features of complex partial seizures in this age group. This has been the object of multiple studies in the past 10–15 years. Holmes [13] analyzed the clinical and EEG features of 69 complex partial seizures in 24 children (average age 12.1 years; range 5–18). The seizures were divided into three electroclinical types: (1) initial motionless stare followed by automatisms and associated with lateralized EEG changes, (2) automatic behavior with bilateral diffuse EEG changes, and (3) loss of body tone followed by automatisms with bilateral EEG changes. The author concluded that complex partial seizures in children are similar to those reported in adults. However, this study did not include infants. Two years later, the same author described 198 partial seizures in 56 children [14] and added to the three types described above a fourth subgroup characterized by initial motor phenomena. Again, no infants were included, since the average age of the patients with complex partial seizures was 13.1 years and the range was 5 to 19 years. Yamamoto et al. [15] analyzed complex partial seizures in 38 children ranging in age from 0 to 13 years. In each child, only the first satisfactorily videotaped seizure was analyzed. They compared 15 children who were two years old or less with 23 children who were 3 to 13 years old. In infants less than three years, the mean seizure duration was substantially longer (120.6 s) than in children who were older (40.6 s). Infants had less purposeful automatisms, which were predominantly oro-alimentary, whereas older children had mostly gestural automatisms. Infants also had more pronounced convulsive movements, usually tonic or myoclonic movements that involved the limbs only or were more widespread. Although all complex partial seizures were said to be associated with disturbance of contact or unresponsiveness, no mention is made of how this was ascertained in infants. Interictal paroxysmal EEG discharges were seen less often in younger children. The location of the initial ictal EEG discharge was variable and did not seem to correlate with the clinical seizure manifestations which were rather uniform regardless of the EEG location. In the same year, Duchowny reported his observations on 187 complex partial seizures recorded in 14 infants under 2 years of age [16]. Their mean age was 8.4 months, with a
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
range of 23 days to 22 months. The most frequent clinical seizure manifestations were behavioral arrest with forced lateralized deviation of head and eyes and extensor stiffening of one upper extremity. Oro-facial automatisms were also common. The issue of consciousness was addressed in this study. Assessment of consciousness during seizures was considered to be difficult. The conclusion that it was at least partially impaired was based on the observation that the infants did not react to visual threat or loud noise during seizures, and that all spontaneous purposeful activities were interrupted. The seizure duration was variable. Seizures lasted less than 1 min in five patients and 1 to 5 min in five patients. Four patients had repetitive seizures referred to as status epilepticus. The interictal EEG recordings were characterized by multifocal epileptiform discharges, generalized slowing, and voltage attenuation, whereas ictal paroxysmal EEG patterns were always lateralized. Watanabe et al. [17] described a syndrome of benign complex partial epilepsies in infancy. Nine infants were described who, between the ages of 3 and 10 months, began to present seizures characterized by motion arrest, decreased responsiveness, staring, simple automatisms, and mild convulsive movements. The interictal EEG was normal and the ictal EEG discharge was either temporal, central, frontal, or parietal. Seizures were easily controlled by carbamazepine or phenobarbital and all patients remained seizure-free for at least 3 years. A similar benign syndrome with partial onset seizures progressing to secondarily generalized was later described [18]. In these patients, the ictal EEG onset was more often central. Watanabe et al. [19] also described a benign idiopathic focal epilepsy in children who had experienced febrile seizures. These children had extratemporal EEG foci and complex partial seizures that were easily controlled, with subsequent remission.
4. Temporal lobe epilepsy in children Both in adults and in children, not all complex partial seizures originate in the temporal lobe. Inversely, it appears that not all seizures originating in the temporal lobe in infants are simple or complex partial seizures. For instance, there have been several isolated reports of infants with infantile spasms who were found to have a temporal lobe tumor [20,21]. The following is a review of our current understanding of temporal lobe epilepsy in infants and children. According to Murakami et al. [22], the incidence of mesial temporal epilepsy in children appears to by quite low, since they found only 19 documented cases among 2319 patients with childhood-onset epilepsy, an incidence of 0.82%. Duchowny et al. [23] described their experience with 16 children less than 12 years old (mean 7 years) who underwent a temporal lobectomy for intractable seizures. After a mean follow-up of 1.5 years (range 1 to 4 years), 11 of the
137
16 children were seizure-free. Seizures consisted mostly of behavioral arrest, stereotyped automatisms, and tonic motor phenomena. Ictal EEG recordings involved the sphenoidal electrode in five of the children; however, initial or maximal activity was never at the sphenoidal electrode but rather in the basal or lateral convexity region of the temporal lobe. The authors concluded that sphenoidal electrodes are less helpful in children than in adults for localizing temporal lobe seizure origin. Pathology in the resected surgical specimens consisted of mesial temporal sclerosis in only two of the 16 patients. In both cases, mesial temporal sclerosis was not an isolated finding. It accompanied focal cortical dysplasia in one case and a neural/glial hamartoma in the other case. The pathological findings were migrational defects, focal cortical dysplasia, low-grade neoplasms, gliosis, tuberous sclerosis, and a hamartoma. Conclusions from these observations are that, in contrast to adults, temporal lobe epilepsy in early childhood is generally neither due to mesial temporal sclerosis nor is it of mesial temporal origin. Wyllie et al. [24] reported their observations in 14 children with temporal lobe epilepsy. This was the first series of children in which the temporal origin of the seizures was clearly established, since only children who became seizurefree for at least 6 months (mean 16 months) after temporal lobectomy were included in the study. The pathology consisted of a tumor in nine of the children, including gangliogliomas, low-grade astrocytomas, and one oligodendroglioma. The location of the tumors was mesiotemporal in five patients, basal temporal in three, and superior lateral temporal in one. One child had cortical dysplasia. The remaining four patients had MRI and PET findings suggestive of mesiotemporal sclerosis but, because resections were not performed en bloc, pathologic confirmation was not possible. Three of these four patients had a history of febrile convulsions as opposed to none of the remaining 10 patients. This difference is statistically significant but does not help to clarify the issue of whether complex febrile seizures cause later mesiotemporal sclerosis or are more likely to occur in infants who are developing mesial temporal sclerosis. The mean age at the time of the video-EEG recordings was 8 years (range 16 months to 12 years). The mean age of onset of the complex partial seizures was 3.8 years (range 6 months to 9 years). Auras were described by nine patients, as young as 5 years and included abdominal/gustatory sensations, somatosensory or visual phenomena, dizziness, or headache. Seizure semiology invariably included decreased responsiveness and automatisms. Gestural and oral automatisms were seen in all but one patient, including in the 16month-old child. Younger patients tended to have less complex gestural automatisms. Other ictal features seen each in a small number of patients included gagging, retching, or vomiting, eye blinking, as well as head nodding or tonic/ clonic motor manifestations at the end of the seizure. Head version and other lateralizing symptoms were rare. EEG findings in three of the four patients with presumed
138
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
mesiotemporal sclerosis, and in the patient with cortical dysplasia, were consistent with those seen also in adults with this condition. The fourth patient with mesiotemporal sclerosis had no interictal spikes. Ictal recordings with scalp electrodes revealed a seizure pattern originating from one hemisphere, whereas invasive recordings localized the seizure pattern to the foramen ovale electrode on the same side. Seven of the nine children with tumors did not have interictal and ictal EEG patterns originating from the affected temporal lobe, but rather extratemporal or falsely lateralized activity. Invasive recordings in three patients did not further clarify the issue. In a similar analysis, Brockhaus and Elger [25] reviewed the clinical manifestations of complex partial seizures of temporal lobe onset in 29 children. After temporal lobectomy, seizures were completely controlled in 24, whereas one had isolated auras and four had a .75% reduction. Their mean age was 11 years, with a range of 18 months to 16 years. Although the patients were initially divided into three age groups (≤6 years, 7 to 12 years and ≥13 years), the final analysis was limited to two age groups: ≤6 years and ≥7 years. Seizure onset was more often out of sleep in the younger children. They were also more likely to have initial motor phenomena, whereas in older children initial auras (epigastric sensation, fearful expression) or psychomotor arrest were more common. The most frequent features of complex partial seizures in the younger children were motor phenomena and simple automatisms followed by versive movements, hypermotoric activity and dystonic postures. No complex automatisms were seen in this age group. The three youngest children had seizures suggestive of infantile spasms. Automatisms became more complex with increasing age, and tonic or clonic symptomatology was observed less frequently. Regarding the clinical symptomatology of seizures of temporal lobe origin in children, the authors conclude that children .6 years old displayed a symptomatology that was similar to temporal lobe seizures of adults. In younger children, the seizures were more likely to be generalized convulsive or to resemble frontal lobe seizures. In this study, ictal EEG recordings in the younger children revealed predominantly generalized discharges without clear focality, whereas in the older children the ictal EEG onset consisted mostly of rhythmic activity over the temporal region. Analysis of the lateralizing value of asymmetric clinical seizure semiology revealed that unilateral clonic activity was always contralateral to the side of ictal EEG onset. Unilateral tonic stiffening was not always contralateral to the EEG seizure onset, and versive movements had little lateralizing value. The localizing value of clinical seizure symptomatology in infants was assessed by Acharya et al. [26]. A total of 125 seizures in 23 infants aged 2 to 24 months were analyzed. The authors were able to delineate two groups based on electro-clinical correlations. One type of seizure consisted primarily of a decrease in behavioral motor activity, with an
indeterminate level of consciousness and minimal or no automatisms. These seizures were characterized as hypomotor and arose from areas behind or below the central fissure, namely the temporal, temporo-parietal, or parieto-occipital regions. Seizures with clonic, tonic, or atonic phenomena, whether localized or bilateral, arose from frontal, frontocentral, central, or frontoparietal areas. Infantile spasms could arise from either location. This finding of a topographic correlation with different seizure symptomatology is in contrast with the observations of Yamamoto et al. [15] who concluded that the location of the initial ictal EEG discharge of complex partial seizures in children was variable and did not seem to correlate with the clinical seizure manifestations which were rather uniform regardless of the EEG location. The main features distinguishing complex partial seizures in infants and young children from those seen in older patients are summarized in Table 1.
5. When does mesial temporal lobe sclerosis develop? Overall, this review of the literature on temporal lobe epilepsy in infants and children provides a fairly homogeneous and concordant picture of the distinguishing electroclinical features of temporal lobe epilepsy in infants and young children, as compared to the syndrome seen in adolescents and adults. However, in terms of underlying etiology, infants and young children with temporal lobe epilepsy represent a much less homogeneous syndrome than adults, and at least three subgroups may be represented. The first group consists of children with focal congenital brain abnormalities or low-grade tumors. A second group may include patients with benign partial epilepsy of infancy [17,18]. Finally, the history of patients with the syndrome of mesial temporal lobe sclerosis reveals at times that the seizure onset occurred during early childhood. Therefore, some of the children with temporal lobe epilepsy either have or will develop mesial temporal sclerosis. This raises the question of when mesial temporal sclerosis develops. Is mesial temporal sclerosis ever present at birth? If not, when and why does it develop? If a child whose MRI shows mesial temporal sclerosis had a previous neuroimaging study without evidence of mesial temporal sclerosis, is Table 1 Clinical characteristics of complex partial seizures in infants, compared with older children and adolescents 1 2 3 4 5
Usually no identifiable aura before the age of 5 years Automatisms predominantly oro-alimentary, less purposeful, less gestural More convulsive activity tonic, versive, clonic, myoclonic Longer duration of seizures Less interictal paroxysmal EEG discharges, ictal discharges less focal
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
it due to improved technology or did the mesial temporal sclerosis develop during the interval between the two tests? The advent of high-resolution MRI may allow to address this question in a prospective manner, but for the time being, this issue can only be assessed by retrospective or crosssectional analyses based on MRI findings or on the pathological analysis of surgical specimens. Mizrahi et al. [27] reported a series of 22 patients with temporal lobe epilepsy of childhood onset who underwent an anterior temporal lobectomy. Their age at the time of surgery ranged from 7 to 36 years. In this series, the two youngest patients with verified Ammon’s horn sclerosis were 13 and 15 years old. In the series by Duchowny et al. [23], 16 children underwent a temporal lobectomy between the ages of 2 and 11 years. Only two had mesial temporal sclerosis: a 10-year-old with focal cortical dysplasia and a 5-year-old with a neural/glial hamartoma. The 14 patients reported by Wyllie et al. [24] underwent a temporal lobectomy between the ages of 6 months and 12 years. Mesial temporal sclerosis was found in four. Their age at surgery was 5, 6, 8, and 9 years, respectively. In a series of 44 children surgically treated for partial epilepsy, Kuzniecky et al. [28] identified nine children with hippocampal sclerosis documented by pathology. Only two of these children were below the age of 12 years; their age is not specified. A cross-sectional MRI study involving 53 children with refractory temporal lobe epilepsy was reported by Grattan-Smith et al. [29]. The mean age was 10 years, with a range of 2–17 years. Hippocampal sclerosis was identified by MRI in 30 of these children (57%). Unfortunately, no breakdown by age is provided. Cross et al. [30] found hippocampal abnormalities in a 4-year-old child, and Harvey et al. [31] identified hippocampal sclerosis in a 2year-old. In another series [25], 29 children underwent a temporal lobectomy between the ages of 18 months and 16 years. Gliosis and/or loss of neurons was described in seven, the youngest being 10 years old. Among 75 children who underwent various types of resection for intractable epilepsy, Polkey [32] found seven cases of mesial temporal sclerosis. There was none in the 0–5-year-old group, one in the 6–11Table 2 Reported cases of mesial temporal sclerosis in infants and children Publication
Youngest patients in series with mesial temporal sclerosis
Mizrahi et al. [27] Duchowny et al. [23] Wyllie at al. [24] Cross et al. [30] Harvey et al. [31] Brockhaus and Elger [25] Murakami et al. [22] Nohria et al. [33] DeLong et al. [34] Gaillard et al. [36]
13 and 15 years 5 and 10 years 5, 6, 8, and 10 years 4 years 2 years 10 years 2 4/12 and 4 3/12 years 2 10/12 years 4, 6, 23, and 30 months 6 children less than 5 years
139
year-old group, and six in the 11–15-year-old group. Murakami et al. [22] reported a series of 19 patients who had mesial temporal lobe epilepsy with onset during childhood. Two patients had MRI evidence of hippocampal atrophy before the age of 5 years, one at the age of 4 years and 3 months, and one at the age of 2 years and 4 months. Nohria et al. [33] reported their very interesting and unique observations in a young child in whom they were able to prospectively follow the development of hippocampal sclerosis. The child was first evaluated by a child neurologist at the age of 16 months because she was delayed in achieving developmental milestones and had a head circumference above the 98th percentile. Her workup was negative, including a normal head MRI scan which was not of high resolution but included coronal sections. At the age of 32 months she presented in status epilepticus involving the left side of her body. Less than 24 h after the seizure she underwent fast spin-echo magnetic resonance imaging with hippocampal volumetry, which revealed an increased T2 signal of the right hippocampus without evidence of hippocampal atrophy. One month later she began to experience complex partial seizures and she had three more episodes of status epilepticus between the ages of 33 and 37 months. She had follow-up MRI studies at 34 and 45 months of age showing progressive hippocampal atrophy and resolution of the increased T2 signal. Because the seizures remained intractable, a right temporal lobectomy was performed when the patient was 51 months old. On pathological examination, the neocortex showed normal lamination with evidence of microdysgenesis. There were increased numbers of neurons in the white matter and rare clusters of large neurons in the deeper layers of the gray matter. Unfortunately, the surgical specimen available for pathological analysis did not include the hippocampus. A syndrome of early-life bilateral hippocampal sclerosis has been recently described [34]. Four infants, 4 months, 6 months, 23 months, and 30 months old, were found to have MRI-documented bilateral atrophy of the hippocampi. The MRI studies were performed at the time of the first clinical seizures, which consisted of afebrile status epilepticus in three and intractable afebrile seizures in one. In addition to the seizures, the clinical picture was that of severe infantile autism. Three of the children had sustained perinatal insults. Two infants underwent a PET study which revealed isolated bilateral anterior temporal lobe hypometabolism. In a very recent report, Harvey et al. [35] described their findings in 63 children with new-onset temporal lobe epilepsy before the age of 15 years. The mean age of onset of the partial seizures was 6.3 years, with a bimodal distribution characterized by peaks in the second and ninth years of life. Abnormal MRI findings were present in 23 of 63 children, with unilateral hippocampal sclerosis in one. Individual ages of patients with hippocampal sclerosis are not provided. Other MRI findings included a tumor in eight patients, one arachnoid cyst, and one case of cortical dysplasia. The authors divided the children into three etiologi-
140
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
cal subgroups: (1) developmental temporal lobe epilepsy with nonprogressive tumors and malformations (10/63), (2) temporal lobe epilepsy with hippocampal sclerosis and/or significant antecedents (18/63), and (3) cryptogenic temporal lobe epilepsy (34/63). The mean ages of seizure onset for these three groups did not differ markedly (8.2 years, 5.1 years, and 6.2 years, respectively, for groups 1, 2 and 3). Most recently, Gaillard et al. [36] reviewed their experience in identifying mesial temporal sclerosis by MRI in children and adolescents. They defined mesial temporal sclerosis as either decreased volume or increased T2 signal of the hippocampal formation. Mesial temporal sclerosis was identified in six children younger than 5 years and in nine children between 5 and 10 years. The authors concluded that mesial temporal sclerosis may be more common in younger children than previously assumed. Table 2 summarizes the above review of reported cases of mesial temporal sclerosis in infants and children documented by MRI and/or pathology. This list contains a total of 17 children with mesial temporal sclerosis in whom this diagnosis was established at the age of 5 years old or less. In conclusion, although there is a growing body of evidence suggesting that mesial temporal sclerosis may develop during early childhood and may be the consequence of prolonged seizures, mesial temporal sclerosis appears to be quite uncommon in younger children, and it still remains unclear why the affected children had prolonged seizures in the first place.
References [1] Wieser HG. Surgically remediable temporal lobe syndromes. In: Engel J Jr, editor. Surgical treatment of the epilepsies, 2nd ed. New York: Raven Press, 1993:49–63. [2] Ebner A, Hoppe M. Noninvasive electroencephalography and mesial temporal sclerosis. J Clin Neurophysiol 1995;12:23–31. [3] King D, Spencer S. Invasive electroencephalography in mesial temporal sclerosis. J Clin Neurophysiol 1995;12:32–45. [4] Cascino GD, Jack CR Jr, Parisi JE, Sharbrough FW, Hirschorn KA, Meyer FB, et al. Magnetic resonance imaging-based volume studies in temporal lobe epilepsy: pathological correlations. Ann Neurol 1991;30:31–36. [5] Henry TR, Chugani HT, Abou-Khalil BW, Theodore WH, Swartz BE. Positron emission tomography. In: Engel J Jr, editor. Surgical treatment of the epilepsies, 2nd ed. New York: Raven Press, 1993:211–232. [6] Arnold S, Schlaug G, Niemann H, Ebner A, Lu¨ders H, Witte OW, et al. Topography of interictal glucose hypometabolism in unilateral mesiotemporal epilepsy. Neurology 1996;46:1422–1430. [7] Henry TR, Babb TL, Engel J Jr, Mazziotta JC, Phelps ME, Crandall PH. Hippocampal neuronal loss and regional hypometabolism in temporal lobe epilepsy. Ann Neurol 1994;36:925–927. [8] Newton MR, Berkovic SF, Austin MC, Reutens DC, McKay WJ, Bladin PF. Dystonia, clinical lateralization and regional cerebral blood flow changes in temporal lobe seizures. Neurology 1992; 42:371–377. [9] Swanson TH. The pathophysiology of human mesial temporal lobe epilepsy. J Clin Neurophysiol 1995;12:2–22.
[10] Piredda S, Gale K. A crucial epileptogenic site in the deep prepiriform cortex. Nature 1985;317:623–625. [11] Babb TL, Brown WJ, Pretorius J, Davenport C, Lieb JP, Crandall PH. Temporal lobe volumetric cell densities in temporal lobe epilepsy. Epilepsia 1984;25:729–740. [12] Sagar HJ, Oxbury JM. Hippocampal neuronal loss in temporal lobe epilepsy: Correlation with early childhood convulsions. Ann Neurol 1987;22:334–340. [13] Holmes GL. Partial complex seizures in children: An analysis of 69 seizures in 24 patients using EEG FM radiotelemetry and videotape recording. Electroenceph clin Neurophysiol 1984;57:13–20. [14] Holmes GL. Partial seizures in children. Pediatrics 1986;77:725– 731. [15] Yamamoto N, Watanabe K, Negoro T, Takaesu E, Aso K, Furune S, et al. Complex partial seizures in children: ictal manifestations and their relationship to clinical course. Neurology 1987;37:1379–1382. [16] Duchowny MS. Complex partial seizures of infancy. Arch Neurol 1987;44:911–914. [17] Watanabe K, Yamamoto N, Negoro T, Takaesu E, Kozaburu A, Furune S, et al. Benign complex partial epilepsies in infancy. Pediatr Neurol 1987;3:208–211. [18] Watanabe K, Negoro T, Aso K. Benign partial epilepsy with secondarily generalized seizures in infancy. Epilepsia 1993;34:635–638. [19] Watanabe K, Takahashi I, Negoro T, Aso K, Miura K. Benign epilepsy of children with complex partial seizures following febrile convulsions. Seizure 1993;2:57–61. [20] Mimaki T, Ono J, Yabuuchi H. Temporal lobe astrocytoma with infantile spasms. Ann Neurol 1983;14:695–696. [21] Asanuma H, Wakai S, Tanaka T, Chiba S. Brain tumors associated with infantile spasms. Pediatr Neurol 1995;12:361–364. [22] Murakami N, Ohno S, Oka E, Tanaka A. Mesial temporal lobe epilepsy in childhood. Epilepsia 1996;37 (suppl 3):52–56. [23] Duchowny M, Levin B, Jayakar P, Resnick T, Alvarez L, Morrison G, et al. Temporal lobectomy in early childhood. Epilepsia 1992; 33:298–303. [24] Wyllie E, Chee M, Granstro¨m ML, DelGiudice E, Estes M, Comair Y, et al. Temporal lobe epilepsy in early childhood. Epilepsia 1993;34:859–868. [25] Brockhaus A, Elger CE. Complex partial seizures of temporal lobe origin in children of different age groups. Epilepsia 1995;36:1173– 1181. [26] Acharya JN, Wyllie E, Lders HO, Kotagal P, Lancman M, Coelho M. Seizure symptomatology in infants with localization-related epilepsy. Neurology 1997;48:189–196. [27] Mizrahi EM, Kellaway P, Grossman RG, Rutecki PA, Armstrong D, Rettig G, et al. Anterior temporal lobectomy and medically refractory temporal lobe epilepsy of childhood. Epilepsia 1990;31:302– 312. [28] Kuzniecky R, Murro A, King D, Morawetz R, Smith J, Powers R, et al. Magnetic resonance imaging in childhood intractable partial epilepsies: pathologic correlations. Neurology 1993;43:681–687. [29] Grattan-Smith JD, Harvey AS, Desmond PM, Chow CW. Hippocampal sclerosis in children with intractable temporal lobe epilepsy: detection with MR imaging. Am J Roentgenol 1993;161:1045–1048. [30] Cross JH, Jackson GD, Neville BGR, Connelly A, Kirkham FJ, Boyd SG, et al. Early detection of abnormalities in partial epilepsy using magnetic resonance. Arch Dis Child 1993;69:104–109. [31] Harvey AS, Grattan-Smith JD, Desmond PM, Chow CW, Berkovic SF. Febrile seizures and hippocampal sclerosis: frequent and related findings in intractable temporal lobe epilepsy of childhood. Pediatr Neurol 1995;12:201–206. [32] Polkey CE. Surgical treatment of epilepsy in children. In: Wallace S, editor. Epilepsy in children. London: Chapman and Hall, 1996:561– 579. [33] Nohria V, Lee N, Tien RD, Heinz ER, Smith JS, DeLong GR, et al. Magnetic resonance imaging evidence of hippocampal sclerosis in progression: a case report. Epilepsia 1994;35:1332–1336.
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141 [34] DeLong GR, Heinz ER. The clinical syndrome of early-life bilateral hippocampal sclerosis. Ann Neurol 1997;42:11–17. [35] Harvey AS, Berkovic SF, Wrennall JA, Hopkins IJ. Temporal lobe epilepsy in childhood: clinical, EEG, and neuroimaging findings and syndrome classification in a cohort with new-onset seizures. Neurology 1997;49:960–968.
141
[36] Gaillard WD, Phulwani P, Dubovsky EC, Conry J, Weinstein S, Vezina LG. Mesial temporal sclerosis in children and adolescents. Epilepsia 1997;38 (suppl 8):137.
Review article
Temporal lobe epilepsy in infants and children Blaise F.D. Bourgeois* Washington University, Department of Neurology, Suite 12E25, St. Louis Children’s Hospital, One Children’s Place, St. Louis, MO 63110, USA Received 24 October 1997; revised version received 6 January 1998; accepted 20 January 1998
Abstract Temporal lobe epilepsy in adults and adolescents is a fairly homogeneous syndrome, both in terms of seizure semiology and in terms of its pathology, and it has been studied extensively. Temporal lobe epilepsy in infants and young children has begun to receive increasing attention in recent years, and a different clinico-pathological picture has emerged. Clinically, the concept of complex partial seizures, which may be useful in adults, is difficult to apply to infants, since it is often not possible to assess impairment of consciousness in this age group. The main distinctive features of complex partial seizures of temporal lobe origin in infants are (1) a predominance of behavioral arrest with possible impairment of consciousness, (2) no identifiable aura, (3) automatisms that are discrete and mostly orofacial, (4) more prominent convulsive activity, and (5) a longer duration (more than 1 min). In addition, seizures of temporal lobe origin in infants may appear clinically generalized, such as infantile spasms or generalized tonic seizures, or can occasionally represent a benign syndrome. The neuropathological findings of temporal lobe epilepsy in infants differ even more than the clinical seizure semiology. In contrast to adult and adolescent patients, mesial temporal sclerosis is a rare finding in infants, in whom the pathological abnormalities associated with seizures of temporal lobe origin consist mostly of dysplasias, migrational disorders, hamartomas, and low-grade tumors such as gangliogliomas. Mesial temporal sclerosis is seen more often in older children than in infants, and its pathogenesis remains a subject of controversy. 1998 Elsevier Science B.V. Keywords: Complex partial seizures; Temporal lobe; Infants; Children
1. Introduction Overall, among the epilepsies of adult and adolescent patients, the temporal lobe plays an extremely important role. In this age group, the temporal lobe is by far the most commonly involved lobe in partial epilepsies, and temporal lobe epilepsy represents a fairly homogeneous syndrome [1]. In the past decade, temporal lobe epilepsy and complex partial seizures in infants and children have been the focus of many clinical investigations and observations. From this work, a picture has emerged that suggests that temporal lobe epilepsy in infants and children represents a different clinico-pathological entity. The present review will focus first on the classical mesial temporal lobe epilepsy of the adult, then the concept of complex partial seizures in infants and children will be discussed
* Fax: +1 314 454 2523; e-mail: [email protected] 0387-7604/98/$19.00 1998 Elsevier Science B.V. All rights reserved PII S0387-7604 (98 )0 0010-2
and, finally, our current understanding of the clinico-pathological aspects of temporal lobe epilepsy in infants and children will be presented.
2. Mesial temporal lobe epilepsy syndrome Among adolescents and adults with the mesial temporal lobe epilepsy syndrome [1], there is an increased incidence of a family history of epilepsy and febrile seizures. A personal history of febrile seizures, which are often prolonged, has been reported in 9–50% of the patients. The onset of the nonfebrile seizures occurs mostly between the ages of 5 and 10 years. They are almost invariably complex partial seizures. Auras, most commonly with epigastric onset and a duration of 5 to 60 s, are experienced by about three-quarters of the patients. The actual complex partial seizure consists of an arrest of activity with a motionless stare, clouding of consciousness, and oroalimentary and hand automatisms.
136
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
There is at times dystonic posturing of the contralateral hand. The duration of the seizure is mostly between 30 and 90 s, with a postictal reorientation that may last several minutes. When spikes are captured on interictal scalp EEG recordings [2], they are usually anterior temporal, with maximum at the sphenoidal electrodes, ear electrodes, or anterior temporal electrodes (T1/T2, FT9/FT10). Contralateral temporal spikes can be seen in 50% of the patients. Ictal surface EEG seizure patterns usually consist of rhythmical activity in the delta, theta, or alpha frequency range, mostly in the ipsilateral temporal region or, less often, over the ipsilateral hemisphere. Invasive ictal EEG recordings reveal 13–25 Hz discharges in medial temporal lobe structures, amygdala and anterior hippocampus [3]. This ictal activity propagates initially to the ipsilateral temporal neocortex in 60%, and to the contralateral hippocampus in 30% of the cases [3]. The characteristic MRI findings of mesial temporal lobe epilepsy represent a triad which consists of (1) hippocampal atrophy on T1-weighted images, (2) an increased hippocampal signal on T2-weighted images, and (3) a loss of internal structure of the hippocampus (decreased white/gray differentiation) [4]. Interictal positron emission tomography (PET) scanning reveals a zone of ipsilateral temporal and extratemporal hypometabolism that is much larger than the EEG zone of interictal activity and ictal onset, and also larger than the zone of neuropathological changes [5]. The zone of hypometabolism involves the temporal lobe in a rather diffuse fashion and often also the ipsilateral hypothalamus. In a more recent study [6], depression of cerebral glucose metabolism was also found in the fronto-orbital cortex bilaterally, as well as ipsilaterally in the posterior insula and in the thalamus. In addition, patients with left temporal lobe epilepsy were found to have hypometabolism in the left inferior frontal gyrus and in the left superior temporal gyrus. No corresponding areas of hypometabolism were seen in patients with right temporal lobe epilepsy. No correlation could be demonstrated between the pattern of hypometabolism on PET scans and the neuronal density of resected hippocampi [7]. Single photon emission computed tomography (SPECT) scanning shows temporal hypoperfusion interictally, but interictal SPECT has a lower resolution than interictal PET scanning. Ictal SPECT scanning reveals temporal hyperperfusion during the seizure; immediately after the seizure, there is mesial temporal hyperperfusion and lateral temporal hypoperfusion [8]. The anatomy and pathophysiology of mesial temporal lobe epilepsy have been studied extensively [9]. Although recordings with invasive electrodes in humans suggest that seizures begin in the hippocampus, the clinical methods are relatively insensitive. It is conceivable that activity generated outside the hippocampus, such as in the pyriform cortex, may become detectable only after propagation and hippocampal amplification [10]. Anatomically, mesial sclerosis is characterized by hippocampal neuronal loss of
at least 30%, and the CA1 region is consistently reported as being the most affected [11]. It is possible that neuronal loss precedes the onset of clinical seizures, but ongoing seizure activity may contribute to the sclerosis. In specimens from 32 temporal lobectomies, there was a highly significant association between severe neuronal loss and early childhood seizures that were prolonged and/or lateralized [12].
3. Complex partial seizures in infants and children Since complex partial seizures are the predominant seizure type in the mesial temporal lobe epilepsy syndrome, it is appropriate to begin a discussion of temporal lobe epilepsy in infants and children with a review of the electroclinical features of complex partial seizures in this age group. This has been the object of multiple studies in the past 10–15 years. Holmes [13] analyzed the clinical and EEG features of 69 complex partial seizures in 24 children (average age 12.1 years; range 5–18). The seizures were divided into three electroclinical types: (1) initial motionless stare followed by automatisms and associated with lateralized EEG changes, (2) automatic behavior with bilateral diffuse EEG changes, and (3) loss of body tone followed by automatisms with bilateral EEG changes. The author concluded that complex partial seizures in children are similar to those reported in adults. However, this study did not include infants. Two years later, the same author described 198 partial seizures in 56 children [14] and added to the three types described above a fourth subgroup characterized by initial motor phenomena. Again, no infants were included, since the average age of the patients with complex partial seizures was 13.1 years and the range was 5 to 19 years. Yamamoto et al. [15] analyzed complex partial seizures in 38 children ranging in age from 0 to 13 years. In each child, only the first satisfactorily videotaped seizure was analyzed. They compared 15 children who were two years old or less with 23 children who were 3 to 13 years old. In infants less than three years, the mean seizure duration was substantially longer (120.6 s) than in children who were older (40.6 s). Infants had less purposeful automatisms, which were predominantly oro-alimentary, whereas older children had mostly gestural automatisms. Infants also had more pronounced convulsive movements, usually tonic or myoclonic movements that involved the limbs only or were more widespread. Although all complex partial seizures were said to be associated with disturbance of contact or unresponsiveness, no mention is made of how this was ascertained in infants. Interictal paroxysmal EEG discharges were seen less often in younger children. The location of the initial ictal EEG discharge was variable and did not seem to correlate with the clinical seizure manifestations which were rather uniform regardless of the EEG location. In the same year, Duchowny reported his observations on 187 complex partial seizures recorded in 14 infants under 2 years of age [16]. Their mean age was 8.4 months, with a
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
range of 23 days to 22 months. The most frequent clinical seizure manifestations were behavioral arrest with forced lateralized deviation of head and eyes and extensor stiffening of one upper extremity. Oro-facial automatisms were also common. The issue of consciousness was addressed in this study. Assessment of consciousness during seizures was considered to be difficult. The conclusion that it was at least partially impaired was based on the observation that the infants did not react to visual threat or loud noise during seizures, and that all spontaneous purposeful activities were interrupted. The seizure duration was variable. Seizures lasted less than 1 min in five patients and 1 to 5 min in five patients. Four patients had repetitive seizures referred to as status epilepticus. The interictal EEG recordings were characterized by multifocal epileptiform discharges, generalized slowing, and voltage attenuation, whereas ictal paroxysmal EEG patterns were always lateralized. Watanabe et al. [17] described a syndrome of benign complex partial epilepsies in infancy. Nine infants were described who, between the ages of 3 and 10 months, began to present seizures characterized by motion arrest, decreased responsiveness, staring, simple automatisms, and mild convulsive movements. The interictal EEG was normal and the ictal EEG discharge was either temporal, central, frontal, or parietal. Seizures were easily controlled by carbamazepine or phenobarbital and all patients remained seizure-free for at least 3 years. A similar benign syndrome with partial onset seizures progressing to secondarily generalized was later described [18]. In these patients, the ictal EEG onset was more often central. Watanabe et al. [19] also described a benign idiopathic focal epilepsy in children who had experienced febrile seizures. These children had extratemporal EEG foci and complex partial seizures that were easily controlled, with subsequent remission.
4. Temporal lobe epilepsy in children Both in adults and in children, not all complex partial seizures originate in the temporal lobe. Inversely, it appears that not all seizures originating in the temporal lobe in infants are simple or complex partial seizures. For instance, there have been several isolated reports of infants with infantile spasms who were found to have a temporal lobe tumor [20,21]. The following is a review of our current understanding of temporal lobe epilepsy in infants and children. According to Murakami et al. [22], the incidence of mesial temporal epilepsy in children appears to by quite low, since they found only 19 documented cases among 2319 patients with childhood-onset epilepsy, an incidence of 0.82%. Duchowny et al. [23] described their experience with 16 children less than 12 years old (mean 7 years) who underwent a temporal lobectomy for intractable seizures. After a mean follow-up of 1.5 years (range 1 to 4 years), 11 of the
137
16 children were seizure-free. Seizures consisted mostly of behavioral arrest, stereotyped automatisms, and tonic motor phenomena. Ictal EEG recordings involved the sphenoidal electrode in five of the children; however, initial or maximal activity was never at the sphenoidal electrode but rather in the basal or lateral convexity region of the temporal lobe. The authors concluded that sphenoidal electrodes are less helpful in children than in adults for localizing temporal lobe seizure origin. Pathology in the resected surgical specimens consisted of mesial temporal sclerosis in only two of the 16 patients. In both cases, mesial temporal sclerosis was not an isolated finding. It accompanied focal cortical dysplasia in one case and a neural/glial hamartoma in the other case. The pathological findings were migrational defects, focal cortical dysplasia, low-grade neoplasms, gliosis, tuberous sclerosis, and a hamartoma. Conclusions from these observations are that, in contrast to adults, temporal lobe epilepsy in early childhood is generally neither due to mesial temporal sclerosis nor is it of mesial temporal origin. Wyllie et al. [24] reported their observations in 14 children with temporal lobe epilepsy. This was the first series of children in which the temporal origin of the seizures was clearly established, since only children who became seizurefree for at least 6 months (mean 16 months) after temporal lobectomy were included in the study. The pathology consisted of a tumor in nine of the children, including gangliogliomas, low-grade astrocytomas, and one oligodendroglioma. The location of the tumors was mesiotemporal in five patients, basal temporal in three, and superior lateral temporal in one. One child had cortical dysplasia. The remaining four patients had MRI and PET findings suggestive of mesiotemporal sclerosis but, because resections were not performed en bloc, pathologic confirmation was not possible. Three of these four patients had a history of febrile convulsions as opposed to none of the remaining 10 patients. This difference is statistically significant but does not help to clarify the issue of whether complex febrile seizures cause later mesiotemporal sclerosis or are more likely to occur in infants who are developing mesial temporal sclerosis. The mean age at the time of the video-EEG recordings was 8 years (range 16 months to 12 years). The mean age of onset of the complex partial seizures was 3.8 years (range 6 months to 9 years). Auras were described by nine patients, as young as 5 years and included abdominal/gustatory sensations, somatosensory or visual phenomena, dizziness, or headache. Seizure semiology invariably included decreased responsiveness and automatisms. Gestural and oral automatisms were seen in all but one patient, including in the 16month-old child. Younger patients tended to have less complex gestural automatisms. Other ictal features seen each in a small number of patients included gagging, retching, or vomiting, eye blinking, as well as head nodding or tonic/ clonic motor manifestations at the end of the seizure. Head version and other lateralizing symptoms were rare. EEG findings in three of the four patients with presumed
138
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
mesiotemporal sclerosis, and in the patient with cortical dysplasia, were consistent with those seen also in adults with this condition. The fourth patient with mesiotemporal sclerosis had no interictal spikes. Ictal recordings with scalp electrodes revealed a seizure pattern originating from one hemisphere, whereas invasive recordings localized the seizure pattern to the foramen ovale electrode on the same side. Seven of the nine children with tumors did not have interictal and ictal EEG patterns originating from the affected temporal lobe, but rather extratemporal or falsely lateralized activity. Invasive recordings in three patients did not further clarify the issue. In a similar analysis, Brockhaus and Elger [25] reviewed the clinical manifestations of complex partial seizures of temporal lobe onset in 29 children. After temporal lobectomy, seizures were completely controlled in 24, whereas one had isolated auras and four had a .75% reduction. Their mean age was 11 years, with a range of 18 months to 16 years. Although the patients were initially divided into three age groups (≤6 years, 7 to 12 years and ≥13 years), the final analysis was limited to two age groups: ≤6 years and ≥7 years. Seizure onset was more often out of sleep in the younger children. They were also more likely to have initial motor phenomena, whereas in older children initial auras (epigastric sensation, fearful expression) or psychomotor arrest were more common. The most frequent features of complex partial seizures in the younger children were motor phenomena and simple automatisms followed by versive movements, hypermotoric activity and dystonic postures. No complex automatisms were seen in this age group. The three youngest children had seizures suggestive of infantile spasms. Automatisms became more complex with increasing age, and tonic or clonic symptomatology was observed less frequently. Regarding the clinical symptomatology of seizures of temporal lobe origin in children, the authors conclude that children .6 years old displayed a symptomatology that was similar to temporal lobe seizures of adults. In younger children, the seizures were more likely to be generalized convulsive or to resemble frontal lobe seizures. In this study, ictal EEG recordings in the younger children revealed predominantly generalized discharges without clear focality, whereas in the older children the ictal EEG onset consisted mostly of rhythmic activity over the temporal region. Analysis of the lateralizing value of asymmetric clinical seizure semiology revealed that unilateral clonic activity was always contralateral to the side of ictal EEG onset. Unilateral tonic stiffening was not always contralateral to the EEG seizure onset, and versive movements had little lateralizing value. The localizing value of clinical seizure symptomatology in infants was assessed by Acharya et al. [26]. A total of 125 seizures in 23 infants aged 2 to 24 months were analyzed. The authors were able to delineate two groups based on electro-clinical correlations. One type of seizure consisted primarily of a decrease in behavioral motor activity, with an
indeterminate level of consciousness and minimal or no automatisms. These seizures were characterized as hypomotor and arose from areas behind or below the central fissure, namely the temporal, temporo-parietal, or parieto-occipital regions. Seizures with clonic, tonic, or atonic phenomena, whether localized or bilateral, arose from frontal, frontocentral, central, or frontoparietal areas. Infantile spasms could arise from either location. This finding of a topographic correlation with different seizure symptomatology is in contrast with the observations of Yamamoto et al. [15] who concluded that the location of the initial ictal EEG discharge of complex partial seizures in children was variable and did not seem to correlate with the clinical seizure manifestations which were rather uniform regardless of the EEG location. The main features distinguishing complex partial seizures in infants and young children from those seen in older patients are summarized in Table 1.
5. When does mesial temporal lobe sclerosis develop? Overall, this review of the literature on temporal lobe epilepsy in infants and children provides a fairly homogeneous and concordant picture of the distinguishing electroclinical features of temporal lobe epilepsy in infants and young children, as compared to the syndrome seen in adolescents and adults. However, in terms of underlying etiology, infants and young children with temporal lobe epilepsy represent a much less homogeneous syndrome than adults, and at least three subgroups may be represented. The first group consists of children with focal congenital brain abnormalities or low-grade tumors. A second group may include patients with benign partial epilepsy of infancy [17,18]. Finally, the history of patients with the syndrome of mesial temporal lobe sclerosis reveals at times that the seizure onset occurred during early childhood. Therefore, some of the children with temporal lobe epilepsy either have or will develop mesial temporal sclerosis. This raises the question of when mesial temporal sclerosis develops. Is mesial temporal sclerosis ever present at birth? If not, when and why does it develop? If a child whose MRI shows mesial temporal sclerosis had a previous neuroimaging study without evidence of mesial temporal sclerosis, is Table 1 Clinical characteristics of complex partial seizures in infants, compared with older children and adolescents 1 2 3 4 5
Usually no identifiable aura before the age of 5 years Automatisms predominantly oro-alimentary, less purposeful, less gestural More convulsive activity tonic, versive, clonic, myoclonic Longer duration of seizures Less interictal paroxysmal EEG discharges, ictal discharges less focal
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
it due to improved technology or did the mesial temporal sclerosis develop during the interval between the two tests? The advent of high-resolution MRI may allow to address this question in a prospective manner, but for the time being, this issue can only be assessed by retrospective or crosssectional analyses based on MRI findings or on the pathological analysis of surgical specimens. Mizrahi et al. [27] reported a series of 22 patients with temporal lobe epilepsy of childhood onset who underwent an anterior temporal lobectomy. Their age at the time of surgery ranged from 7 to 36 years. In this series, the two youngest patients with verified Ammon’s horn sclerosis were 13 and 15 years old. In the series by Duchowny et al. [23], 16 children underwent a temporal lobectomy between the ages of 2 and 11 years. Only two had mesial temporal sclerosis: a 10-year-old with focal cortical dysplasia and a 5-year-old with a neural/glial hamartoma. The 14 patients reported by Wyllie et al. [24] underwent a temporal lobectomy between the ages of 6 months and 12 years. Mesial temporal sclerosis was found in four. Their age at surgery was 5, 6, 8, and 9 years, respectively. In a series of 44 children surgically treated for partial epilepsy, Kuzniecky et al. [28] identified nine children with hippocampal sclerosis documented by pathology. Only two of these children were below the age of 12 years; their age is not specified. A cross-sectional MRI study involving 53 children with refractory temporal lobe epilepsy was reported by Grattan-Smith et al. [29]. The mean age was 10 years, with a range of 2–17 years. Hippocampal sclerosis was identified by MRI in 30 of these children (57%). Unfortunately, no breakdown by age is provided. Cross et al. [30] found hippocampal abnormalities in a 4-year-old child, and Harvey et al. [31] identified hippocampal sclerosis in a 2year-old. In another series [25], 29 children underwent a temporal lobectomy between the ages of 18 months and 16 years. Gliosis and/or loss of neurons was described in seven, the youngest being 10 years old. Among 75 children who underwent various types of resection for intractable epilepsy, Polkey [32] found seven cases of mesial temporal sclerosis. There was none in the 0–5-year-old group, one in the 6–11Table 2 Reported cases of mesial temporal sclerosis in infants and children Publication
Youngest patients in series with mesial temporal sclerosis
Mizrahi et al. [27] Duchowny et al. [23] Wyllie at al. [24] Cross et al. [30] Harvey et al. [31] Brockhaus and Elger [25] Murakami et al. [22] Nohria et al. [33] DeLong et al. [34] Gaillard et al. [36]
13 and 15 years 5 and 10 years 5, 6, 8, and 10 years 4 years 2 years 10 years 2 4/12 and 4 3/12 years 2 10/12 years 4, 6, 23, and 30 months 6 children less than 5 years
139
year-old group, and six in the 11–15-year-old group. Murakami et al. [22] reported a series of 19 patients who had mesial temporal lobe epilepsy with onset during childhood. Two patients had MRI evidence of hippocampal atrophy before the age of 5 years, one at the age of 4 years and 3 months, and one at the age of 2 years and 4 months. Nohria et al. [33] reported their very interesting and unique observations in a young child in whom they were able to prospectively follow the development of hippocampal sclerosis. The child was first evaluated by a child neurologist at the age of 16 months because she was delayed in achieving developmental milestones and had a head circumference above the 98th percentile. Her workup was negative, including a normal head MRI scan which was not of high resolution but included coronal sections. At the age of 32 months she presented in status epilepticus involving the left side of her body. Less than 24 h after the seizure she underwent fast spin-echo magnetic resonance imaging with hippocampal volumetry, which revealed an increased T2 signal of the right hippocampus without evidence of hippocampal atrophy. One month later she began to experience complex partial seizures and she had three more episodes of status epilepticus between the ages of 33 and 37 months. She had follow-up MRI studies at 34 and 45 months of age showing progressive hippocampal atrophy and resolution of the increased T2 signal. Because the seizures remained intractable, a right temporal lobectomy was performed when the patient was 51 months old. On pathological examination, the neocortex showed normal lamination with evidence of microdysgenesis. There were increased numbers of neurons in the white matter and rare clusters of large neurons in the deeper layers of the gray matter. Unfortunately, the surgical specimen available for pathological analysis did not include the hippocampus. A syndrome of early-life bilateral hippocampal sclerosis has been recently described [34]. Four infants, 4 months, 6 months, 23 months, and 30 months old, were found to have MRI-documented bilateral atrophy of the hippocampi. The MRI studies were performed at the time of the first clinical seizures, which consisted of afebrile status epilepticus in three and intractable afebrile seizures in one. In addition to the seizures, the clinical picture was that of severe infantile autism. Three of the children had sustained perinatal insults. Two infants underwent a PET study which revealed isolated bilateral anterior temporal lobe hypometabolism. In a very recent report, Harvey et al. [35] described their findings in 63 children with new-onset temporal lobe epilepsy before the age of 15 years. The mean age of onset of the partial seizures was 6.3 years, with a bimodal distribution characterized by peaks in the second and ninth years of life. Abnormal MRI findings were present in 23 of 63 children, with unilateral hippocampal sclerosis in one. Individual ages of patients with hippocampal sclerosis are not provided. Other MRI findings included a tumor in eight patients, one arachnoid cyst, and one case of cortical dysplasia. The authors divided the children into three etiologi-
140
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141
cal subgroups: (1) developmental temporal lobe epilepsy with nonprogressive tumors and malformations (10/63), (2) temporal lobe epilepsy with hippocampal sclerosis and/or significant antecedents (18/63), and (3) cryptogenic temporal lobe epilepsy (34/63). The mean ages of seizure onset for these three groups did not differ markedly (8.2 years, 5.1 years, and 6.2 years, respectively, for groups 1, 2 and 3). Most recently, Gaillard et al. [36] reviewed their experience in identifying mesial temporal sclerosis by MRI in children and adolescents. They defined mesial temporal sclerosis as either decreased volume or increased T2 signal of the hippocampal formation. Mesial temporal sclerosis was identified in six children younger than 5 years and in nine children between 5 and 10 years. The authors concluded that mesial temporal sclerosis may be more common in younger children than previously assumed. Table 2 summarizes the above review of reported cases of mesial temporal sclerosis in infants and children documented by MRI and/or pathology. This list contains a total of 17 children with mesial temporal sclerosis in whom this diagnosis was established at the age of 5 years old or less. In conclusion, although there is a growing body of evidence suggesting that mesial temporal sclerosis may develop during early childhood and may be the consequence of prolonged seizures, mesial temporal sclerosis appears to be quite uncommon in younger children, and it still remains unclear why the affected children had prolonged seizures in the first place.
References [1] Wieser HG. Surgically remediable temporal lobe syndromes. In: Engel J Jr, editor. Surgical treatment of the epilepsies, 2nd ed. New York: Raven Press, 1993:49–63. [2] Ebner A, Hoppe M. Noninvasive electroencephalography and mesial temporal sclerosis. J Clin Neurophysiol 1995;12:23–31. [3] King D, Spencer S. Invasive electroencephalography in mesial temporal sclerosis. J Clin Neurophysiol 1995;12:32–45. [4] Cascino GD, Jack CR Jr, Parisi JE, Sharbrough FW, Hirschorn KA, Meyer FB, et al. Magnetic resonance imaging-based volume studies in temporal lobe epilepsy: pathological correlations. Ann Neurol 1991;30:31–36. [5] Henry TR, Chugani HT, Abou-Khalil BW, Theodore WH, Swartz BE. Positron emission tomography. In: Engel J Jr, editor. Surgical treatment of the epilepsies, 2nd ed. New York: Raven Press, 1993:211–232. [6] Arnold S, Schlaug G, Niemann H, Ebner A, Lu¨ders H, Witte OW, et al. Topography of interictal glucose hypometabolism in unilateral mesiotemporal epilepsy. Neurology 1996;46:1422–1430. [7] Henry TR, Babb TL, Engel J Jr, Mazziotta JC, Phelps ME, Crandall PH. Hippocampal neuronal loss and regional hypometabolism in temporal lobe epilepsy. Ann Neurol 1994;36:925–927. [8] Newton MR, Berkovic SF, Austin MC, Reutens DC, McKay WJ, Bladin PF. Dystonia, clinical lateralization and regional cerebral blood flow changes in temporal lobe seizures. Neurology 1992; 42:371–377. [9] Swanson TH. The pathophysiology of human mesial temporal lobe epilepsy. J Clin Neurophysiol 1995;12:2–22.
[10] Piredda S, Gale K. A crucial epileptogenic site in the deep prepiriform cortex. Nature 1985;317:623–625. [11] Babb TL, Brown WJ, Pretorius J, Davenport C, Lieb JP, Crandall PH. Temporal lobe volumetric cell densities in temporal lobe epilepsy. Epilepsia 1984;25:729–740. [12] Sagar HJ, Oxbury JM. Hippocampal neuronal loss in temporal lobe epilepsy: Correlation with early childhood convulsions. Ann Neurol 1987;22:334–340. [13] Holmes GL. Partial complex seizures in children: An analysis of 69 seizures in 24 patients using EEG FM radiotelemetry and videotape recording. Electroenceph clin Neurophysiol 1984;57:13–20. [14] Holmes GL. Partial seizures in children. Pediatrics 1986;77:725– 731. [15] Yamamoto N, Watanabe K, Negoro T, Takaesu E, Aso K, Furune S, et al. Complex partial seizures in children: ictal manifestations and their relationship to clinical course. Neurology 1987;37:1379–1382. [16] Duchowny MS. Complex partial seizures of infancy. Arch Neurol 1987;44:911–914. [17] Watanabe K, Yamamoto N, Negoro T, Takaesu E, Kozaburu A, Furune S, et al. Benign complex partial epilepsies in infancy. Pediatr Neurol 1987;3:208–211. [18] Watanabe K, Negoro T, Aso K. Benign partial epilepsy with secondarily generalized seizures in infancy. Epilepsia 1993;34:635–638. [19] Watanabe K, Takahashi I, Negoro T, Aso K, Miura K. Benign epilepsy of children with complex partial seizures following febrile convulsions. Seizure 1993;2:57–61. [20] Mimaki T, Ono J, Yabuuchi H. Temporal lobe astrocytoma with infantile spasms. Ann Neurol 1983;14:695–696. [21] Asanuma H, Wakai S, Tanaka T, Chiba S. Brain tumors associated with infantile spasms. Pediatr Neurol 1995;12:361–364. [22] Murakami N, Ohno S, Oka E, Tanaka A. Mesial temporal lobe epilepsy in childhood. Epilepsia 1996;37 (suppl 3):52–56. [23] Duchowny M, Levin B, Jayakar P, Resnick T, Alvarez L, Morrison G, et al. Temporal lobectomy in early childhood. Epilepsia 1992; 33:298–303. [24] Wyllie E, Chee M, Granstro¨m ML, DelGiudice E, Estes M, Comair Y, et al. Temporal lobe epilepsy in early childhood. Epilepsia 1993;34:859–868. [25] Brockhaus A, Elger CE. Complex partial seizures of temporal lobe origin in children of different age groups. Epilepsia 1995;36:1173– 1181. [26] Acharya JN, Wyllie E, Lders HO, Kotagal P, Lancman M, Coelho M. Seizure symptomatology in infants with localization-related epilepsy. Neurology 1997;48:189–196. [27] Mizrahi EM, Kellaway P, Grossman RG, Rutecki PA, Armstrong D, Rettig G, et al. Anterior temporal lobectomy and medically refractory temporal lobe epilepsy of childhood. Epilepsia 1990;31:302– 312. [28] Kuzniecky R, Murro A, King D, Morawetz R, Smith J, Powers R, et al. Magnetic resonance imaging in childhood intractable partial epilepsies: pathologic correlations. Neurology 1993;43:681–687. [29] Grattan-Smith JD, Harvey AS, Desmond PM, Chow CW. Hippocampal sclerosis in children with intractable temporal lobe epilepsy: detection with MR imaging. Am J Roentgenol 1993;161:1045–1048. [30] Cross JH, Jackson GD, Neville BGR, Connelly A, Kirkham FJ, Boyd SG, et al. Early detection of abnormalities in partial epilepsy using magnetic resonance. Arch Dis Child 1993;69:104–109. [31] Harvey AS, Grattan-Smith JD, Desmond PM, Chow CW, Berkovic SF. Febrile seizures and hippocampal sclerosis: frequent and related findings in intractable temporal lobe epilepsy of childhood. Pediatr Neurol 1995;12:201–206. [32] Polkey CE. Surgical treatment of epilepsy in children. In: Wallace S, editor. Epilepsy in children. London: Chapman and Hall, 1996:561– 579. [33] Nohria V, Lee N, Tien RD, Heinz ER, Smith JS, DeLong GR, et al. Magnetic resonance imaging evidence of hippocampal sclerosis in progression: a case report. Epilepsia 1994;35:1332–1336.
B.F.D. Bourgeois / Brain & Development 20 (1998) 135–141 [34] DeLong GR, Heinz ER. The clinical syndrome of early-life bilateral hippocampal sclerosis. Ann Neurol 1997;42:11–17. [35] Harvey AS, Berkovic SF, Wrennall JA, Hopkins IJ. Temporal lobe epilepsy in childhood: clinical, EEG, and neuroimaging findings and syndrome classification in a cohort with new-onset seizures. Neurology 1997;49:960–968.
141
[36] Gaillard WD, Phulwani P, Dubovsky EC, Conry J, Weinstein S, Vezina LG. Mesial temporal sclerosis in children and adolescents. Epilepsia 1997;38 (suppl 8):137.