If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Bitemporal epilepsy (BTLE) significantly differs from unitemporal epilepsy (UTLE).
•
BTLE patients had an older age at epilepsy onset and rare positive brain MRI.
•
Interictal/ictal EEG activity and dystonia were more frequently bilateral in BTLE.
•
Head deviation and oroalimentary automatisms were less frequent in BTLE.
•
The identified pattern allows the non-invasive recognition of BTLE.
Abstract
Purpose
Temporal lobe epilepsy (TLE) with bilateral ictal involvement (bitemporal epilepsy, BTLE) is an intriguing form of TLE whose characteristics need to be carefully identified as BTLE patients are not good surgical candidates. The purpose of this study was to define the anatomo-electro-clinical features differentiating BTLE from unilateral TLE (UTLE).
Methods
Forty-eight BTLE patients underwent long-term video-EEG monitoring (VEEG) and experienced seizures with bilateral temporal lobe involvement. Their main electro-clinical (demographics, interictal and ictal EEG, ictal signs) and neuro-imaging [brain magnetic resonance imaging (MRI)] data were compared with those of a group of 38 UTLE patients.
Results
In comparison with the UTLE patients, the BTLE cohort was significantly older at the time of epilepsy onset (p = 0.023), more frequently experienced bilateral asynchronous interictal epileptiform discharges during wakefulness (p = 0.001) and sleep (p < 0.001), bilateral upper limb dystonia (p = 0.005), and auditory auras (p = 0.027), and less frequently showed a recognisable initial ictal EEG pattern of focal flattening or low-voltage fast activity (p < 0.001), post-ictal memory of seizures (p = 0.001), staring (p < 0.001), head deviation (p = 0.004), oro-alimentary automatisms (p = 0.006), and positive brain MRI (p < 0.001). MRI revealed neoplastic lesions (p = 0.007) or alterations other than hippocampal sclerosis (p = 0.028) only in the UTLE patients.
Conclusion
The possibility of recognising BTLE patients during pre-surgical evaluation or being able to suspect bitemporal seizures before VEEG by identifying particular anatomo-electro-clinical patterns is diagnostically important for epileptologists and can help to prevent possible surgical failures.
Temporal lobe epilepsy (TLE) is the most frequently observed drug-resistant localisation-related epilepsy, especially in epilepsy surgery case series,[
]. Before proceeding to epilepsy surgery, it is essential to evaluate the localisation of the epileptogenic zone (EZ), which can be assessed by collecting a careful history and undertaking detailed anatomo-electro-clinical investigations [
]. If a patient's history, seizure semiology, interictal and ictal electroencephalographic (EEG) findings, and magnetic resonance imaging (MRI) and positron emission tomography (PET) neuro-imaging data are consistent with a diagnosis of unilateral TLE (UTLE), he or she can be considered a good candidate for surgery [
TLE patients showing bilateral temporal lobe involvement during seizures may not be optimal surgical candidates. The existence of bitemporal epilepsy (BTLE) is well known and many studies have considered the possible use of surgical therapeutic strategies, albeit with conflicting results [
]. However, BTLE has not yet been clearly defined and relatively little is known about the anatomo-electro-clinical data characterising BTLE patients. Hirsch et al. [
] failed to find any significant clinical or neuroradiological differences between BTLE and UTLE patients, although the former had a less frequent history of febrile seizures (FS). Schulz et al. [
] found an association between BTLE and the absence of auras, and it has been recently reported that BTLE is associated with bilateral lateralising ictal signs [
]. It has also been found that BTLE patients less frequently show ictal motor signs than those with UTLE, and experience longer periods of post-ictal unresponsiveness [
On the basis of the assumption that pathophysiological mechanisms and epileptic networks are different in BTLE and UTLE, it is possible that the two forms have different anatomo-electro-clinical features. The aim of this study was to investigate the electro-clinical and neuro-imaging characteristics of a cohort of BTLE patients in an attempt to differentiate them from UTLE patients because the identification of an anatomo-electro-clinical phenotype of BTLE would help clinical epileptologists to recognise or at least suspect BTLE when evaluating TLE patients, especially in the case of a pre-surgical assessment.
2. Materials and methods
2.1 Data collection
This multicentre study involved three tertiary referral centres for epilepsy and epilepsy surgery in Milan, Italy: the Claudio Munari Epilepsy Surgery Centre of Niguarda Hospital, the Clinical Epileptology Unit of Carlo Besta Neurological Institute, and the Epilepsy Centre of San Paolo Hospital. Although belonging to different hospitals, these centres cooperate in the context of a multidisciplinary Inter-Hospital Department for Pre-surgical Epilepsy Evaluation (DDEP).
We retrospectively reviewed the electronic charts of 2152 epilepsy patients admitted to the three centres for pre-surgical evaluations between January 1995 and March 2013: 1548 underwent epilepsy surgery, and 48 (2.2%) were identified as having drug-resistant BTLE. The control group consisted of 38 consecutive patients with recorded seizures who underwent surgery because of refractory UTLE between 2005 and 2011 and had been seizure-free for at least 24 months.
Each of the patients in the two cohorts had undergone long-term scalp video-EEG (VEEG) using a digital VEEG recording device (Nihon-Kohden Neurofax or Micromed System Plus Evolution) and electrodes placed according to the international 10-20 system. A total of 223 seizures were recorded in the BTLE group and 86 in the UTLE group. Four BTLE patients also underwent long-term intra-cerebral electrode stereo-EEG (SEEG) monitoring (Nihon-Kohden Neurofax). The patients’ anti-epileptic drug (AED) doses were usually reduced by at least 50% in order to facilitate the recording of seizures.
All of the patients underwent high-resolution MRI performed using a 1.5 T (Siemens Avanto or Philips ACS-NT & Achieva) or 3 T (Philips Achieva TX) scanner. The MRI protocol included transverse spin-echo double-echo images of the entire brain, coronal fast spin-echo T2-weighted and spin-echo FLAIR images, and coronal fast spin-echo inversion recovery T1-weighted images. The transversal and coronal sections were respectively acquired in parallel with or perpendicularly to the axis of the hippocampal formation.
BTLE was diagnosed on the basis of two ictal electro-clinical inclusion criteria: a VEEG and/or SEEG recording of at least one seizure simultaneously or sequentially involving the two temporal lobes, without the possibility of lateralising its onset and subsequent development [defined as a non-lateralisable (NL) bitemporal seizure], and/or the recording of at least two different seizures alternately arising from the two temporal lobes [defined as an independent (IND) bitemporal seizure]. More precisely, a bitemporal seizure was considered NL in the presence of a discharge simultaneously involving the two temporal lobes (‘synchronous’) or spreading from one temporal lobe to the opposite side and back (‘asynchronous’), or ending on the opposite side (‘side-switching’) (Figs. 1 and 2A and B ).
Fig. 1Ictal EEG activity in non-lateralisable BTLE. Two MRI-negative patients showing a synchronous ictal onset and asynchronous ictal discharge propagation. (A) Diffuse attenuation can be seen for three seconds from the point marked by the asterisk, followed by bitemporal non-lateralisable (NL) theta-delta rhythmic activity (circle), which accelerates in the right temporal region after 14 s and ends at the same time in the two temporal rows of electrodes. (B) The arrow indicates the beginning of bitemporal NL theta-delta rhythmic activity, which accelerates in the right temporal region seven seconds after onset and then in the left temporal region, 20 seconds later, ending about two seconds before the end of the discharge in the right temporal electrodes. BTLE: bitemporal epilepsy; MRI: magnetic resonance imaging.
Fig. 2Ictal EEG activity in independent BTLE and UTLE. (A) Right temporal EEG activity changes from the point marked by the asterisk, with low-amplitude delta activity evolving into rhythmic theta activity in the same region. (B) Another seizure in the same patient showing initial low-amplitude theta activity in the left temporal region (circle) that evolves into rhythmic delta activity in the same region. (C) Coronal FLAIR MRI recording of the same patient showing left HS; (D) Coronal FLAIR MRI recording of a UTLE patient showing right HS. (E) EEG recording of the same UTLE patient during a seizure showing low-voltage fast ictal activity on the anterior and intermediate temporal leads (rectangle) that evolves into rhythmic spikes in the same region. BTLE: bitemporal epilepsy; UTLE: unitemporal epilepsy; MRI: magnetic resonance imaging; HS: hippocampal sclerosis.
The BTLE patients were divided into two groups (NL or IND BTLE) on the basis of a blinded review of each seizure. In the case of disagreement, or when more than one seizure type was recorded in the same patient, the main seizure pattern was agreed after a collegial re-evaluation during which the epileptologists not only assessed the ictal EEG traces of each patient, but also interictal activity by visually analysing at least ten 20-second wakeful and sleeping interictal traces samples.
After distinguishing the BTLE and UTLE patients, we collected all of their available clinical, neurophysiological and neuro-imaging data, concentrating on the main historical information, the results of a neurological examination, ictal semiology, ictal and interictal EEG findings, and MRI and, when available, fluorodeoxyglucose PET (FDG-PET) data.
2.2 Statistical analysis
Contingency table analysis was used to evaluate the associations between the different types of epilepsy and the nominal or dichotomous variables, with the independence of the rows and columns being tested by means of Fisher's exact test. Age at the time of epilepsy onset, age at the time of evaluation/surgery, disease duration and monthly seizure frequency were compared in the UTLE and two BTLE subgroups using the Kruskal–Wallis test. A p-value of <0.05 was considered significant.
The significant variables in the three groups underwent post hoc analysis using a Bonferroni correction for multiple comparisons, with a p-value of <0.017 being considered significant.
The data were analysed using SPSS software (version 22).
3. Results
3.1 General patient characteristics
A total of 48 BTLE patients were identified between January 1995 and March 2013; the control group consisted of 38 consecutive seizure-free patients who underwent surgery for UTLE between 2005 and 2011. Table 1 shows their general characteristics.
The difference in the proportion of BTLE and UTLE patients with positive family history of epilepsy (8.3% vs 28.9%) was of borderline significance (p = 0.051). There were no significant differences in personal antecedents or the presence of FS between the two groups except in the case of central nervous system (CNS) infections, which were reported in only five BTLE patients (10.4%) (p = 0.018). The UTLE patients were younger at the time of epilepsy onset (p = 0.023).
There were no significant between-group differences in terms of gender distribution, age at the time of evaluation/surgery, disease duration, monthly seizure frequency, the circadian rhythm of the seizures, the presence of a >1-year seizure-free period, or the presence of secondary generalised seizures.
None of the patients in either group had any significant neurological examination abnormality.
3.2 Electro-clinical and neuro-imaging findings
3.2.1 Interictal EEG findings
There was no difference in background activity between the BTLE and UTLE patients, but synchronous (and more often asynchronous) bilateral interictal theta-delta slow activity (ISA) (wakefulness, p < 0.001; sleep, p < 0.001), and bilateral interictal epileptiform discharges (IEDs: i.e. spikes, spike-waves and sharp waves) (wakefulness, p = 0.001; sleep, p < 0.001), were significantly more frequent in the temporal regions of the BTLE patients (Table 2 and Fig. 3).
Fig. 3Interictal EEG in BTLE. Asynchronous bilateral interictal epileptiform discharges in the temporal regions during wakefulness (A) and sleep (B) in a patient with non-lateralisable BTLE. BTLE: bitemporal epilepsy.
The BTLE patients showed two different ictal discharge patterns: 32 (66.7%) NL and 16 (33.3%) IND (Fig. 1, Fig. 2A and B).
3.2.3 Initial ictal EEG pattern
Initial ictal EEG flattening or low-voltage fast activity (LVFA) was significantly more frequently observed in the temporal regions of the UTLE patients (p < 0.001); the first detectable ictal EEG modification in the majority of BTLE patients was rhythmic delta or theta activity (Fig. 2A, B and E).
3.2.4 Ictal signs
Seizure semiology was significantly different between the two groups: post-ictal memory of seizures (PiMS) (p = 0.001), staring (p < 0.001), head deviation (HD) (p = 0.004), and oro-alimentary automatisms (OAs) (p = 0.006) were significantly less frequent in the BTLE patients, and PiMS and HD were much less frequent in the NL than in the IND BTLE subgroup. Furthermore, in the latter, the HD could be ipsilateral or contralateral to the ictal discharge, whereas it was ipsilateral in most of the UTLE patients (Tab. 2).
Seizure warning (verbally or by means of pushing a warning button) was significantly more frequent among the IND BTLE than the NL BTLE or UTLE patients (p = 0.004).
The occurrence of auras (including epigastric auras), loss of consciousness (LOC) and gestural automatisms was not significantly different between the BTLE and UTLE patients. However, although not statistically significant, an epigastric aura was present in 52.6% of the UTLE and only 29.2% of the BTLE patients, whereas auditory auras were significantly more frequent in the IND BTLE patients (p = 0.027).
There was no significant difference in upper limb dystonia between the BTLE and UTLE patients, but bilateral dystonia (BD) was only observed in the NL BTLE subgroup and affected 18.8% of the patients (p = 0.005). When present, dystonia was unilateral in the majority of IND BTLE patients (appearing contralaterally to the ictal discharge in most cases), and in all of the UTLE patients in whom it was always contralateral to the ictal discharge (Table 2).
The median duration of VEEG was significantly longer in the BTLE than in the UTLE patients, and longer in those with IND BTLE than in those with NL BTLE (p < 0.001). Furthermore, electro-clinically different seizures were only recorded in the BTLE group, and mainly in the patients with IND BTLE (p < 0.001).
3.2.5 Neuro-imaging findings
Table 2 shows the main MRI findings, and Fig. 2 details of individual cases. MRI was positive in only 20/48 BTLE patients (41.7%), but in 35/38 UTLE patients (92.1%) (p < 0.001), who were the only patients with alterations other than hippocampal sclerosis (HS), especially tumours (p = 0.007); bilateral MRI abnormalities (bilateral HS) were only observed in the BTLE group (in 10.4% of the patients), but this difference was not statistically significant.
Table 2Electro-clinical and neuro-imaging findings.
FDG-PET data were available for 15 BTLE patients (not shown), and revealed normal brain metabolism in five, unilateral temporal hypometabolism in nine, and bilateral temporal hypometabolism in one; these results often disagreed with the MRI findings.
3.2.6 Invasive recordings
Four of the 48 BTLE patients underwent bitemporal SEEG. IND bitemporal seizures were recorded in three, in one of whom they were also synchronous; the fourth patient experienced side-switching bitemporal seizures, which were also recorded during scalp VEEG. None of the four patients underwent surgery.
3.3 Treatments and outcomes
At the time of evaluation, all of the BTLE patients were receiving pharmacological treatment (many with two or more AEDs) and were considered drug resistant [
Only four of the 48 BTLE patients underwent surgery (a right antero-mesial temporal resection). In three cases, pre-surgical VEEG showed that the ictal discharge started in the right temporal lobe, but later also strongly involved the left temporal lobe; they were therefore retrospectively classified as having NL BTLE. MRI revealed right HS in one case, but nothing of note in the other two. In the fourth case, pre-surgical VEEG revealed only right temporal lobe seizures, but contralateral independent seizures were recorded after surgery. MRI was negative.
Histopathology revealed type IIIa focal cortical dysplasia (FCD) in one case, and type I FCD in another; the findings were unremarkable in the remaining two cases.
All of the surgically treated BTLE patients had an unsatisfactory seizure outcome (Engel Class III).
4. Discussion
Despite limitations due to its retrospective nature, this study allowed the identification of an anatomo-electro-clinical pattern that was peculiar to BTLE and distinct from UTLE. In comparison with UTLE patients, the patients with BTLE were older at the time of epilepsy onset, more frequently had auditory auras and bilateral asynchronous interictal abnormalities (particularly during sleep), and less frequently had recognisable initial ictal EEG activity (focal LVFA or flattening), the post-ictal memory of seizures or positive MRI findings, and less frequently showed staring, head deviation or oro-alimentary automatisms. They were also the only patients with bilateral dystonia and bilateral MRI abnormalities (HS).
Most of our results are considerably different from those of previous studies, which found that BTLE patients were less likely to have a history of FS, auras, unitemporal IEDs, ictal motor signs or lateralised mesial temporal MRI pathology, but more likely to show bilateral lateralising ictal signs and to experience a longer period of post-ictal unresponsiveness,[
] particularly those with independent bitemporal seizures.
There were significant differences between our BTLE and UTLE cohorts in terms of both ictal and interictal EEG activity. Our BTLE patients showed more frequent bilateral asynchronous interictal abnormalities during wakefulness and sleeping, which may be a marker of bilateral temporal lobe involvement in the epileptic network (Fig. 3). These abnormalities (especially bilateral asynchronous IEDs) have been previously reported in other studies of BTLE,[
]. Consequently, they cannot be considered sufficient to distinguish the two groups of patients.
Another important difference is the fact that the BTLE patients had a less recognisable initial ictal EEG discharge: i.e. focal flattening or LVFA preceding a rhythmic ictal discharge (Fig. 2). To the best of our knowledge, this has never been reported before, and is probably an expression of a scarcely lateralised epileptic network in BTLE or a very localised discharge in both mesial structures. This plausibly causes a poorly regionalised change in brain electrical activity during seizures and the consequent absence of the early ictal scalp EEG focal signs characterising UTLE patients [
Our BTLE patients showed two different forms of ictal discharge onset and development, which led us to divide them into those with NL BTLE (66.7%) and those with IND BTLE (33.3%): other authors have also adopted a similar sub-categorisation [
]. The fact that AEDs were reduced or withdrawn in order to facilitate seizure recording should not have had any causal influence on the ictal pattern because the same protocol was used for BTLE and UTLE patients, and previous studies have excluded the possibility that reduced AED intake leads to misleading changes in seizure lateralisation or electrical ictal onset during VEEG [
]. We therefore believe that such a sub-categorisation is a reliable and useful means of describing BTLE.
The two BTLE subgroups were significantly different from each other, as well as significantly different from the UTLE group, particularly in relation to some important ictal signs. The patients with NL BTLE recalled their seizures post-ictally and showed ictal HD more rarely than those with IND BTLE or UTLE, and were the only ones to show BD: taken together, these signs indicate probable bilateral temporal lobe ictal involvement. There is both direct and indirect evidence showing that post-ictal memory (and therefore post-ictal recall of an aura and/or seizure) can be impaired if a bitemporal seizure has induced transient bilateral hippocampal dysfunction [
]. The absence of clear-cut interlobar electrical asymmetry during NL bitemporal seizures may lead to more infrequent HD and the presence of BD. It is well known that unilateral dystonia is an important lateralising ictal sign in TLE [
], and we found that it was contralateral to ictal onset in all our UTLE patients and most of the IND BTLE patients presenting it.
Apart from the fact that electro-clinically different seizures were only recorded in our IND BTLE patients (a probable reason for their more frequent warnings of seizures because ictal semiology may also differ in terms of the duration of an aura before LOC), we did not observe any ictal or post-ictal features specifically characterising this subgroup, which can resemble the group of UTLE patients. Nevertheless, the two BTLE sub-groups had some common anatomo-electro-clinical characteristics that differentiated them as a whole from the UTLE patients, in addition to the interictal and initial ictal EEG patterns discussed above.
], who found that the absence of an aura was characteristic of BTLE, we did not find any significant differences between our BTLE and UTLE patients in terms of the presence or absence of an aura (including epigastric auras), although the former more frequently reported auditory auras. Taken together with the less frequent presence of staring and OAs, this previously unreported finding suggests that BTLE patients may have different brain networks from those of UTLE patients, possibly leading to more frequent neocortical than mesial seizure onset and discharge propagation [
] but we found that there was no difference in their distribution in our groups. However, in comparison with our UTLE patients, those with BTLE were characterised by: (1) an infrequent family history of epilepsy; (2) an older age at epilepsy onset; (3) less frequent positive brain MRI scans; (4) no MRI-detected brain tumours; (5) bilateral MRI alterations (HS); and (6) a history of CNS infections (Table 1, Table 2). Except for points 1 and 2, which have never been reported before, these findings are in agreement with those of earlier studies [
]. Moreover, histopathology showed that only two of our four surgically treated BTLE patients had FCD, and other studies have found that the main histological finding in BTLE is HS [
], and have only infrequently described FCD or other focal lesions. All of these data indicate that the etiologies commonly observed in UTLE patients are less probable, or at least differently expressed, in patients with BTLE, thus leading to a distinct anatomo-electro-clinical phenotype.
Bitemporal epileptogenicity cannot be attributed to a longer disease duration or more frequent seizures, as suggested by other authors [
], because there was no difference in these variables between our UTLE and BTLE cohorts. Moreover, as HS seems to be associated with the duration of epilepsy (particularly when FCD is not detected) [
], the comparable duration of epilepsy in the two groups may explain the similar incidence of HS.
The VEEG recordings of our BTLE patients (especially those with IND BTLE) were significantly longer than those of our UTLE patients (Table 2). As there was no significant difference in the median number of recorded seizures per day in the two groups, it is likely that bitemporal epileptogenicity was suspected on the grounds of the patients’ histories. However, the fact that the seizures of some IND BTLE patients may cluster on only one temporal lobe for months [
] may make even the longest VEEG monitoring period insufficient to lateralise a patient's EZ unequivocally. We therefore suggest that IND BTLE should be considered whenever patients report ictal signs that are very different from those documented by VEEG, and that the diagnostic process should be guided by the patient's overall clinical history and ictal signs (including those indicated by anamnestic semiology) and neuro-imaging findings whenever VEEG does not exclude it.
We could not include PET data in the statistical analysis because only some of our patients underwent a PET examination, but we believe that such data might further contribute to identifying BTLE patients, particularly in the presence of a negative or bilateral PET signal alteration. However, this will need to be verified in a specifically designed comparative study.
Previous studies have attempted to identify possible prognostic factors indicating a favourable surgical outcome in patients identified as having BTLE by means of scalp VEEG or invasive recordings [
], but their results are partially contradictory and sometimes based on the percentage reduction in seizure frequency. Some have indicated good surgical outcomes [
]. On the basis of the data relating to our few patients who underwent surgery, patients with BTLE are not good candidates for resection, and probably account for a large proportion of poor TLE surgery outcomes. Other investigators have described the usefulness of invasive recordings [
], but we believe that a careful pre-surgical non-invasive work-up can adequately identify patients with BTLE and that therapeutic strategies other than resective surgery should be further investigated for those who are drug resistant [
], who concluded that BTLE and UTLE are part of a continuum and share common pathophysiological mechanisms, and that UTLE could substantially be considered a bilateral disease, we think that they should be treated as two distinct entities in the context of TLE spectrum, even if they may have some similar characteristics. This is strongly supported by their different anatomo-electro-clinical phenotypes, although it has to be pointed out that these may not apply to patients with non-drug-resistant BTLE or UTLE.
Further studies are still required to investigate other possible different etiologies (channel diseases, genetic causes, SNC infections, auto-immune pathologies) and characterise the neuropsychological and brain metabolism aspects of the disease. Furthermore, our retrospectively identified anatomo-electro-clinical patterns of BTLE might also need to be verified in prospective studies of pre-surgical TLE patients, which could also provide additional electro-clinical and etiopathogenetic information that could lead to new pharmacological or palliative surgical options. Another possibility is to use VEEG in supposedly UTLE patients with bad surgical outcome in order to identify the real origin of the ictal discharges.
5. Conclusions
The findings of this study provide evidence that BTLE and UTLE have different characteristics and may belong to separate symptomatological constellations. They also suggest a particular anatomo-electro-clinical pattern that allows the non-invasive recognition of BTLE and could be used in the pre-surgical evaluation of patients with presumed TLE in order to verify or exclude the presence of possible BTLE.
Conflict of interest
None.
Acknowledgements
We would like to thank Dr. Angelo Del Sole and Dr. Antonio Scarale (specialists in Nuclear Medicine) for the PET scan examinations.
We would also like to thank the Paolo Zorzi Association for the Neurosciences for supporting the Carlo Besta Epilepsy Surgery Group.
00186-7&id=gr1.jpg){kind=link}
00186-7&id=gr2.jpg){kind=link}
00186-7&id=gr3.jpg){kind=link}