JLE

Epileptic Disorders

MENU

Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification Volume 23, issue 4, August 2021

Videos

  • Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification
  • Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification
  • Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification
  • Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification
  • Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification
  • Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification
  • Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification
  • Clinical semiology of temporal lobe seizures in preschool children: contribution of invasive recording to anatomical classification

Figures


  • Figure 1

  • Figure 3

  • Figure 2

  • Figure 4

  • Figure 5

  • Figure 6

  • Figure 7

  • Figure 8

Tables

Temporal lobe epilepsy (TLE) is frequently encountered in children, and TLE patients are generally considered good candidates for epilepsy surgery. Temporal lobe semiology in children has been widely studied based on video-EEG recording material, and compared to TLE in adults in terms of semiology [1-4] and value of lateralizing signs [5]. Nevertheless, electro-clinical semiology of TLE in the subgroup of preschool children has been, to date, poorly evaluated, and rarely on the basis of anatomical classification provided by invasive recordings. We report our findings of a series of preschool children who entered our programme of presurgical evaluation, focusing on clinical and EEG semiology according to the anatomical origin of seizures within the temporal lobe.

Patients and methods

Between 2010 and 2016, we identified 89 consecutive patients from our database who had a surgical resection preceded by a pre-surgical investigation for drug-resistant TLE. We selected a preschool patient subgroup, aged six years and younger, at the time of surgery in order to study the electro-clinical semiology of temporal lobe seizures at a young age. A comprehensive epileptic history was obtained for each patient.

The accurate topography of the seizure onset was assessed by a surface video-EEG recording completed for the majority of patients with an invasive recording. The decision to operate and the modalities of surgery were decided by a multidisciplinary team based on the criteria we use for all patients in this age range [6].

Scalp video-EEGs were performed according to the guidelines published for presurgical video-EEG monitoring [7, 8]. The 10-20 system with 19 electrodes can be used from the age of three months. For TLE, additional temporo-basal electrodes are of great value in detecting interictal spikes and assessing the seizure onset [9, 10]. When possible, a clinical examination was performed during the seizures.

When scalp EEG was not sufficient to determine the seizure onset, invasive explorations were required. A foramen ovale electrode (FOE), combined with scalp electrodes, was very helpful when there were strong arguments for a temporo-mesial epilepsy, in particular to detect subtle seizures [11, 12]. As for other localizations, stereoelectroencephalography (SEEG) was indicated when the anatomical-electroclinical correlations obtained by non-invasive and/or FOE investigations were insufficiently concordant or inconclusive regarding the localization of the supposed epileptic zone [13]. Based on a previous report from out tertiary centre, the ictal onset zone matched the lesion in only 32% of patients with symptomatic epilepsies of this age group [14].

After the age of two, SEEG is the most appropriate invasive method for exploration of temporal epilepsies. The involvement of Wernicke's area, which could be a specific indication for subdural recording, is an exceptional issue in preschool children. All the seizures recorded were reviewed by two neurophysiologists (DT, MC).

We classified the epilepsy according to the precise origin of the seizures within the temporal lobe and followed the anatomical classification proposed by Maillard et al.[15]:

  • mesial structures (hippocampus amygdala and para hippocampal structures);
  • temporal pole;
  • neocortical temporal structures (lateral and basal) (T1 to T3 excluding the pole);
  • and mesio-temporal and lateral regions.

We excluded multilobar epilepsies even if the temporal lobe was part of the epileptogenic zone. These encompass so-called Temporal plus epilepsies, as reported by Barba et al. [16],defined as epilepsies with a typical temporal semiology but a seizure onset zone which includes extratemporal zones, mainly insular and sometimes frontal or parietal areas (six of our patients). These also include multilobar epilepsies defined by onset zones in at least two lobes, either temporal or extratemporal; the seizure semiology is concordant with the corresponding epileptogenic zone (as in seven of our patients).

Pathological studies of the surgical material were analysed in one laboratory (Lariboisiere Hospital). Results were classified according to the International League Against Epilepsy (ILAE) classification for dysplasia [17] and classification of tumours [18].

The outcome of epilepsy was assessed using Engel's classification [19].

Results

Twenty-eight patients (nine female and 19 male), aged six years and under, were studied (table 1 table 1). The mean age at epilepsy onset was 12.8 months (range: 0-54); the first seizure consisted of infantile spasms (IS) in three, complex febrile seizures in eight,and focal seizures in 17 patients. Altogether, three patients presented with some IS during the course of epilepsy that were controlled with medication at the time of the surgery in all cases but one (Patient 22). None of the patients had electrophysiological features of hypsarrhythmia. All patients had associated focal seizures with the exception of one (Patient 22) who had late-onset IS as the only seizure type following herpetic encephalitis. On MRI, a structural lesion was found in all the patients but one (Patient 19).

All patients underwent tailored temporal surgery at a mean age of 3.6 years (range: 0.5 to 6.4). Five patients (18%) were under two years of age at the time of the surgery. Before surgery, all the children underwent a surface video-EEG recording. Twenty-three patients (82%) were explored with invasive electrodes that consisted of FOE (DIXI or ALCIS electrodes) in 11 (39%) and SEEG (DIXI or ALCIS electrodes) in 12 (43%). Two patients underwent both FOE exploration and SEEG because the former led to the hypothesis of a temporo-mesial epilepsy being discarded (table 2 table 2). In thenine patients for whom FOE was contributory, both a temporo-mesial onset for some of the seizures as well as the occurrence of subtle seizures were confirmed.

The patients were classified into four anatomo-electroclinical groups, and for each group, the clinical and EEG data are provided (table 3 table 3), as well as one or more video-EEG samples of a “typical” seizure.

Mesio-temporal seizures (15 patients)

Of these patients, 27% were explored with SEEG. We distinguished three subgroups according to ictal semiology that we defined as follows:

  • Typical mesial seizures are very similar to those in adults and are characterized by the following signs:
  • dysautonomic signs such as face rubefaction, pallor or cyanosis, mydriasis;
  • oro-alimentary signs such as nausea, chewing, salivation, vomiting;
  • emotional and subjective signssuch as fear, epigastric sensation;
  • contact changes with staring;
  • motor signs such as automatisms, typically homolateral to the seizure onset, and posturing manifestation, often contralateral to the epileptic focus.

A typical temporo-mesial case is shown in Patient 3 (video sequence 1 and figure 1).

A typical amygdala seizure with fear is shown in Patient 15 (video sequence 2 and figure 2).

  • (b) Seizures with apnoea. Three patients presented this type of seizure in which the dominant sign was apnoea with deep desaturation requiring oxygenotherapy at home (Patient 1) (video sequence 3, figure 3).
  • (c) Subtle seizures. These are often overlooked and identified based on video-EEG. During sleep, patients may present with a somewhat “odd” awakening. During wakefulness, there are subtle changes in behaviour compared to the child's baseline activity. This category includes hypomotor seizures characterized by behavioural arrest (Patient 5) (video sequence 4, figure 4).

Temporal pole seizures (two patients)

These are subtle seizures, and may thus be easily overlooked. In the video provided for this type of seizure, the seizures were identified based on SEEG and the clinical signs consisted of a hypomotor seizure with speech arrest, motor arrest, and automatisms (Patient 16) (video sequence 5, figure 5). More symptomatic seizures are due to propagation to adjacent structures; an example of a patient presenting with a hyperkinetic seizure mimicking frontal lobe semiology is provided (Patient 17) (video sequence 6, figure 6).

Neocortical temporal seizures (five patients)

In these seizures, there is a lack of “typical” semiology which makes the diagnosis difficult in the absence of depth electrodes. In our patients, the diagnosis was based on SEEG. The clinical signs frequently observed consisted of asymmetric tonic posturing and/or subtle seizures. In the case of a dominant hemisphere, post-ictal aphasia can be present (Patient 19) (video sequence 7, figure 7).

Mesial and neocortical seizures (six patients)

These seizures have a temporo-mesial onset with latero-temporal propagation. After the typical mesial semiology, contralateral motor manifestations are systematically observed, related to suprasylvian propagation via the temporal neocortex. In the case of a dominant hemisphere, in speaking patients, post-ictal aphasia can be seen (Patient 26) (video sequence 8, figure 8).

Post-ictal contralateral paresis was observed in 10 patients and could not be linked to a specific topography within the temporal lobe. More often, the deficit was found during clinical examination or careful review of the video, but rarely reported by the parents. It has a high lateralizing value. Post-ictal aphasia was impossible to assess in 12 patients (43%), either because the patients were too young or had a speech delay (and not cooperative post-ictally), or because of tiredness. Three patients (11%) had clear-cut post-ictal aphasia. Post-ictal language was normal in 11 children (39%) (table 2).

Surgery

Thirty surgical procedures were performed on twenty-eightpatients: two patients were operated on twice (Patients 2 and 7) because of an incomplete mesial resection. The description of the surgical resection for each patient is reported in table 3.

With a mean follow-up duration of 5.5 years (range: 1.8-9.7)and according to Engel's classification, 89% (25)of the patients were classified as Engel Class I, whereas three were Engel Class III (Patients 7, 10 and 22). The reasons for failure were incomplete temporal resection in Patients 10 and 22, while a mutation in the FGF12 gene in Patient 7 could explain the unexpected post-surgical failure.

Pathological studies revealed focal cortical dysplasia (FCD) in 13, an isolated tumour in 10,and a tumourassociated with a FCD in two. Isolated HS was found in one patient, encephalitis in one and histology was negative in one patient. HS was found in 11 patients (one isolated, two associated with a tumour, and eightassociated with FCD).

According to the ILAE pathological classification, FCD I was found in onepatient, FCD II in four, FCD IIIa in eight, and FCD IIIb in two. Isolated tumours were found in 10 patients including seven gangliogliomas, twodysembryoplastic neuroepithelial tumours and one choroid plexus papilloma associated with HScorresponding to the so-called “long-term epilepsy-associated tumour” [18].

Discussion

To our knowledge, this is the first series reporting temporal lobe electro-clinical semiology in preschool children, including a large number of toddlers (19%). Moreover, it is the first report in which clinical semiology is based on electrophysiological findings and not on the topography of the lesion based on MRI [4] or on seizure freedom after surgery [1-3, 20]. Indeed, epilepsy onset is not always determined by the lesion itself, which can be missing or does not match with the epileptic focus, and can arise from remote areas (perilesional onset in the case of a tumour or tuberous sclerosis complex, with difficulties in assessing the onset of seizures in the case of dual pathology).

Epilepsy freedom after surgery cannot always be considered an accurate “a posteriori” localizing element for two reasons:

  • the extent of the resection can go beyond the epileptic focus;
  • in patients with neocortical epilepsy, mesial structures are frequently involved (and vice versa), but their implication may result from spreading rather than a primary seizure onset, thus the entire resection is not always required to obtain seizure freedom.

FOEwas indicated to detect subtle seizures in patients with a temporal lesion, developmental or behavioural delay, and a few apparent seizures [12], but was never indicated to rule out a bitemporal epilepsy, which exceptionally occurs in children. SEEG was carried out in 43% of patients, allowing for greater precision of topography regarding the seizure onset, and thus based on reliable clinical semiology, as previously described [15]. In the literature, invasive exploration is undertaken in up to 50% of children with lesional epilepsies, irrespective of the type [21]. For temporal epilepsies, temporal lobe epilepsies are invasively explored in 7-27% of cases [22-24]. We have already pointed out that we tend to do more invasive explorations than most teams [25].

Specific semiology in children

The paediatric literature addresses temporal lobe semiology in general and focuses on changes in semiology with age. For most authors [2, 3], some clinical signs change with age (mainly motor manifestations), whereas others (neurovegetative features notably) are present at all ages; TLE semiology is very similar to that of adults after the age of six years [1]. In the literature, it appears thatclinical signs frequently reported in preschool children are dominated by motor manifestations (IS, tonic and clonic seizures), described as either symmetrical [3] or asymmetrical [4]. Motor manifestations were not observed as frequently in our series. The first reason for this is that all authors pool together mesial seizures (unrelated to tonic or clonic seizures) with neocortical seizures (typically associated with motor manifestations that occur early). The second reason is that most authors include IS, which are frequent, as part of the motor manifestations. In our patients, three presented with IS, preceded by or associated with focal seizures. In a large series of localization-related IS, we did not find any specific correlation between IS and the temporal lobe, compared to other cortical areas [26]. Tonic-clonic generalization is infrequent in the literature [1-4, 20], as in our patients, with the exception of one child in whom a FGF12 mutation, a possible aetiological factor, was found. Most authors underline the frequent occurrence of hypomotor seizures at this age [2, 3, 20, 27], as in our patients. Our series emphasizes the difficulty in making a diagnosis due to the frequency of so-called “subtle” seizures that are frequently overlooked and often discovered by looking carefully at the video recording and sometimes only by using depth electrode recordings. As for so-called “behavioural seizures”, subtle and hypomotor seizures can be misdiagnosed as non-epileptic behavioural disorders especially in children with psychiatric comorbidity and/or intellectual disability, both frequently associated with epilepsy in general and TLE in particular [28]. This type of subtle seizure has previously been reported in preschool children irrespective of topography [29]. We have also previously reported that these subtle seizures may arise from the prefrontal lobe, making their localization value fairly poor in the absence of electrophysiological findings [30]. Automatisms are less complex in preschool children, compared to older children. They are mainly oral, or subtle bimanual [2-4]. It is also notable that lateralizing ictal and post-ictal features are more frequent in older children [2]. In our experience, post-ictal paresis is rarely reported by the parents, but is found in 24% of patients during the video-EEG recording. Post-ictal aphasia is much more difficult to assess because of the basic speech status of the patient and post-ictal lack of cooperation or because of sleepiness after the seizure. In 22 children, however, a speech assessment was possible, and in only six, aphasia was found. The clinical signs that are reported to be unchanged relative to age are neurovegetative signs, emotional manifestations and auras [2, 3]. However, Fontana et al. underlined the rarity of auras in children younger than six years of age [27]. Regarding emotional signs frequently reported in TLE, we only observed negative emotions such as fear. None of our patients presented with positive emotions, unlike the 100 children reported by Fogarasi et al. [31]. For these authors, emotions have a localizing value for the temporal lobe in 26% and positive emotions have a lateralising value in the right hemisphere. The difference relative to our results is probably due to the younger age of our children. This is in accordance with our observation that auras are rare in infants and young children during focal seizures, irrespective of the localization [25].

Bitemporal epilepsies, which are significant in adult patients, are rarely encountered in children and were notably not encountered in any our patients [32].

Anatomical correlations

In adult patients, a topographic classification of TL seizures has been proposed [15]. Our series shows that, even in preschool patients, there are clinical seizure characteristics that support the hypothesis of precise topography, and that this classification can be used.

Mesial structures

Topographic classification of TL seizures is particularly relevant to mesio-temporal epilepsies. We have shown that, even in toddlers, the semiology of mesio-temporal epilepsy can be very similar to that in adults [33-35], except for subjective signs that are missing in infants and pre-language children. In contrast, children can express some subjective signs, such as fear or pain, at as young as two or three years old, as shown in video sequence 2.

The occurrence of seizures with apnoea and desaturation as a dominant and/or isolated manifestation in infants is not very widely reported in adults but has been observed in children [29, 36]. Apnoea has been reported to be frequently associated with hypomotor seizures [29] and is significantly more frequent in young children with TLE [36]. More precisely, apnoea can appear as a predominant sign in children with mesio-temporal epilepsies. Three patients from our series had this type of seizure and the severity of the desaturation required oxygenotherapy at home. Regarding aetiology, there is a strong link with the existence of tumours, mainly ganglioglioma (two patients in our series and three previously reported patients)[37, 38]. Regarding the anatomical location for apnoea, the amygdala and the head of the hippocampus are clearly involved. Stimulation of these structures in adult patients, investigated using SEEG for epilepsy, can induce apnoea [39] and apnoea can be responsible for sudden unexpected death in epilepsy (SUDEP) or near missed SUDEP [40] in adult patients with complex partial seizures. These seizures, even when rare, must be identified and treated early due to the risk of infant apnoea syndrome.

Pole structures

Temporo-polar epilepsies are not considered to have very specific semiology [41]. The most striking reported feature, compared to mesial epilepsies, is the early loss of consciousness. In our patients, seizures limited to the temporal pole were characterized by a behavioural modification, while other ictal features were related to propagation. In these cases, the apparent semiology was challenging to interpret as it results from a spread to other structures. One example is provided by our patient who presented with a hyperkinetic seizure; on SEEG, the seizure began within the temporal pole and then spread to the cingulum via the anterior hippocampus. Hyperkinetic seizures are classically related to the frontal lobe with different onset zones, such as the orbito-frontal or frontal mesial structures [42]. Their occurrence in TLE has previously been reported in adult patients [43, 44], but to our knowledge never in young children, except for the report by Brokhaus and Elger [1].

Neocortical structures

Children with temporal lateral seizures have non-specific semiology. This is also the case in adults. In the larger series in which semiology was objectively analysed, the initial loss of contact was significantly more frequent in temporal lateral than temporal mesial seizures [15]. The other statistically significant features were early unilateral clonic movements, a shorter duration, and the occurrence of tonic-clonic generalization [15, 45].The most characteristic feature of temporal lateral epilepsies is the auditory aura which significantly differentiates them from mesial epilepsies [15]. Auditory illusions or hallucinations are clearly only reported exceptionally by preschool children.

Conclusion

This series illustrates that in young children with TLE, fairly good anatomical-electroclinical correlations are possible, which contrasts with established opinions [29, 46]. The predominance of motor manifestations reported in young children was not found in our series, probably because previous authors included IS and mesial- and lateral-onset epilepsies were pooled together. Temporal lobe semiology, mainly mesio-temporal, has some similarities with the semiology in adults and can be identified using electro-clinical data; the particularities in young children are the occurrence of overlooked subtle seizures and the occurrence of seizures with apnoea which have a clear mesio-temporal origin and strong link with so-called “long-term epilepsy-associated tumour”.

Supplementary material

Summary slides accompanying the manuscript are available at www.epilepticdisorders.com.

Acknowledgements and disclosures

We thank the physicians who referred the patients for epilepsy surgery, Pr Veronique Sébille (Nantes University) for statistical analysis, our EEG technologists Mrs M. Benghezal and Mrs I. Cheramy, and Mrs. Aisha Yu, a native English speaker, for careful revision of the manuscript.

None of the authors have any conflicts of interest to disclose. None of the authors received any financial support.