Extratemporal ictal clinical features in hippocampal sclerosis: Their relationship to the degree of hippocampal volume loss and to the outcome of temporal lobectomy
Address correspondence to Paolo Borelli, Neurology Department, Università degli studi di Verona, P.le L.A. Scuro 10–37134 Verona, Italy. E-mail: email@example.com
Purpose: Since extratemporal clinical features in patients with unilateral hippocampal sclerosis (HS) are likely to indicate aberrant ictal spread or a more extensive epileptogenic zone, we asked whether such features are associated with more severe HS and a worse outcome following temporal lobectomy.
Patients and methods: We reviewed all patients (174) who had undergone temporal lobectomy for histologically proven unilateral HS related temporal lobe epilepsy between 1997–2005 atthe National Hospital for Neurology and Neurosurgery. We divided patients into those with severe HS (side-to-side ratio < 0.6) and those with mild HS (side-to-side ratio > 0.75). We examined all seizures recorded on electroencephalography (EEG) video telemetry in these patients for clinical features of temporal lobe epilepsy. The postsurgical outcome was classified using the Engel classification at the time of follow up (median 4.7 years, range 1–9 years).
Results: Patients (28 out 39) with severe HS had atypical features compared to 7 out of 27 in the mild HS [Chi square (χ2) test, p = 0.0013]. The mean number of atypical clinical features was 2.2 in the severe HS group and 0.62 in the mild HS group (Mann Whitney U Test, p < 0.001). The percentage of postsurgery seizure freedom (class 1 Engel classification) was 87%, and there was no significant effect of the presence of atypical clinical features.
Conclusions: This study shows that atypical (extratemporal) clinical features tend to occur more frequently in patients with severe HS and do not correlate with worse surgical outcome.
Temporal lobe epilepsy (TLE) is the most common localization-related form of epilepsy, accounting for more than 60%–70% of focal epilepsies (Wieser & ILAE commission on neurosurgery of epilepsy, 2004). Hippocampal sclerosis (HS) is the most common pathology underlying TLE and is detected by magnetic resonance imaging (MRI) in approximately 65% of cases (Blumcke et al., 2002). The typical neuroradiological feature of HS is loss of volume of the affected hippocampus. This is usually associated with T2 signal hyperintensity, while other features such as ipsilateral ventricular enlargement and blurring of the gray-white matter boundary are more variable (Salmenpera & Duncan, 2005). The degree of hippocampal volume loss can be measured with a high degree of accuracy by MRI volumetry (Cascino et al., 1991; Cook et al., 1992).
The classical clinical features of a fully developed complex partial seizure of mesial temporal lobe origin comprises three elements: (1) aura, which commonly takes the form of autonomic, psychic, perceptual, or dysamnestic phenomena; (2) pseudo absence, which takes the form of motor arrest, impairment of awareness and responsiveness, and a blank staring appearance; and (3) automatism, which is typically oroalimentary (i.e., lip smacking or chewing) and/or manual (i.e., “picking up” movements, fidgeting, or handling objects) (Engel, 2001; Wieser & Commission on Neurosurgery of Epilepsy, 2004). Common signs that are reliably lateralizing during the seizure include contralateral spasm and posturing of the upper limb, postictal dysphasia (dominant), and ictal speech (nondominant). Other periictal signs are less common or less reliable (Commission on Classification and Terminology of the International League Against Epilepsy, 1989; Loddenkemper & Kotagal, 2005). The seizure duration is typically less than 2 min in duration, but the seizure is often followed by confusion and disorientation postictally, which resolves gradually over a period of minutes (Commission on Classification and Terminology of the International League Against Epilepsy, 1989; Engel, 2001; Wieser & Commission on Neurosurgery of Epilepsy, 2004).
Seizures of extratemporal cortex origin have clinical features that are generally very different from those seizures originating in the mesial temporal lobe, although there is considerable overlap. The most common ones, grouped into frontal, parietal, occipital, and temporo-parietal-occipital cortex are listed in Table 1 (Commission on Classification and Terminology of the International League Against Epilepsy, 1989; Sveinbjornsdottir & Duncan, 1993; Salanova et al., 1995; Kotagal & Arunkumar, 1998; So, 1998; Siegel & Williamson, 2000; Williamson & Jobst, 2000; Kellinghause & Luders, 2004).
Table 1. Atypical clinical features examined in this study
| • Asymmetric bilateral tonic posturing.|
| • Speech arrest or loud vocalization.|
| • Ictal facial grimacing.|
| • Complex partial seizures with prominent motor behavior (i.e., pedaling, cycling, boxing, rocking movements) or agitated motor behavior.|
| • Prominent head and eyes tonic deviation.|
| • Clonic or tonic–clonic contralateral movements.|
| • Behavior with a strong emotional or aggressive quality (i.e., crying, laughing).|
| • Sudden onset and offset of ictal motor features.|
| • Aura with cephalic, body, or throat sensation.|
| • Aura with forced thinking.|
| • Elementary sensory disturbance (i.e., tingling, numbness) or pain sensations.|
| • Body image disturbance.|
| • Forced blinking, clonic eyeball movements.|
| • Ictal hemianopia.|
| • Simple visual hallucination.|
|Temporoparietooccipital junction symptoms|
| • Multimodal hallucinations.|
Cluster analysis has been used to differentiate the semiology of temporal and extratemporal seizures (Manford et al., 1996). Although effective in determining commonly occurring differences, these analyses do not gauge the importance of rarer clinical features.
Although the symptoms and signs described above are characteristic of seizures arising in different anatomical locations, there is often quite significant overlap due presumably to propagation of seizure discharges from one location to another or to the distributed nature of many seizure networks. This is well illustrated by a recent description of nocturnal hyperkinetic seizures, suggestive of frontal lobe epilepsy, originating in the insular cortex (Ryvlin et al., 2006). The so-called mesial temporal lobe seizures have often widely distributed epileptogenic activity that can be recorded in limbic, temporal neocortical, and frontal cortical areas (Gloor, 1997). This is the case even if surgical resection of the hippocampus results in seizure remission.
The occurrence, in mesial temporal seizures, of those ictal clinical features (signs and symptoms) that are more characteristic of extratemporal origin, has not been systematically reviewed to our knowledge. Carreno et al. (2005) and others reported that complex motor behaviors in temporal lobe epilepsy occurred in only 12 of 502 patients. Furthermore, there have been three cases of sleep-related, hyperkinetic seizures of temporal lobe origin studied with intracranial electroencephalography (EEG) that demonstrated seizure origin in the temporal lobe but rapidly spread to extratemporal structures, commensurate with the onset of the hyperkinetic phase (Nobili et al., 2004). These reports have led to the suggestion that extratemporal semiological features are secondary to spread of the ictus beyond the temporal lobe to the cingulate, frontal, and parietal cortex.
One of the authors (M.C.W.) had the clinical impression that extratemporal features (symptoms and signs) were more common in patients with radiologically severe HS than in patients with radiologically mild HS (for reasons outlined in the discussion section of this paper). If confirmed, this would be an important point when investigating patients for potential temporal lobe surgery, in whom atypical seizure semiology might be mistakenly thought to render surgery inappropriate. Also, it might also be postulated that extratemporal clinical features in a mesial temporal lobe seizure signified a widespread neuronal network that might reduce the chance of seizure freedom after temporal lobe surgery (Ryvlin et al., 2006).
This present study therefore was conceived to test two hypotheses: (1) in patients with HS being assessed for temporal lobe surgery, those with severe HS were more likely to have ictal clinical features, which are generally considered more typical of extratemporal epilepsy, than are those associated with mild HS and (2) the outcome of temporal lobectomy in patients in whom preoperative seizures had extratemporal features was less successful.
We carried out a retrospective study of all patients who (1) had undergone standard anterior temporal lobe resection (the procedure was the same for all patients) and who had a histologically proven diagnosis of HS between 1997–2005; (2) had measurement of hippocampal volumes from volumetric MRI; and (3) had seizures recorded on video EEG telemetry at the National Hospital (174 patients). The year 1997 was chosen as the starting year since this was when measurements of hippocampal volumes from volumetric MRI were introduced into routine presurgical practice at the National Hospital for Neurology and Neurosurgery. Hippocampal volumes were measured using our standard method (Cook et al., 1992; Free et al., 1995) on high-quality volumetric scans. All scans were reviewed by a neuroradiologist (J.M.S.), and we excluded patients with radiological, histological, or historical evidence of dual pathology and patients with neuroradiological evidence of bilateral HS. We did not measure the volumes of other structures (i.e., temporal neocortex, entorhinal cortex, or thalamus) and therefore cannot be certain that there was not subtle volume loss outside the hippocampus. We used the volume ratio, because we have found that absolute volumes can differ depending upon MRI methods used, and over this time period, we changed from using a 1.5 Tesla MRI to a 3 Tesla MRI. To simplify the statistical analysis, we divided the patients into two polar groups based upon the volumetric MRI hippocampal measurements (ratio of volumes of affected to unaffected side). We selected values that best approximated the 20th and 80th percentile. Based on this, we defined two groups: (1) severe HS, defined as patients whose hippocampal ratio was equal or below 0.6 (39 patients) and (2) mild HS, defined as patients in whom the hippocampal ratio was equal or above 0.75 (27 patients). The 107 patients with ratios between 0.6 and 0.75 were not considered further.
The clinical notes of each one of these patients were also reviewed to collect information about their past history. Sixty-three of the 66 patients had interictal and ictal scalp EEG features consistent with mesial temporal lobe epilepsy due to HS. In the remaining three patients, the scalp EEG was nonlocalizing or discordant, and in these three patients invasive EEG recording was undertaken which revealed seizure onset in the sclerotic mesial temporal lobe.
We compiled, based on the published literature (Commission on Classification and Terminology of the International League Against Epilepsy, 1989; Sveinbjsdottir & Duncan, 1993; Salanova et al., 1995; Kotagal & Arunkumar, 1998; So, 1998; Kutsy, 1999; Siegel & Williamson, 2000; Kotagal et al., 2003; Kellinghause & Luders, 2004) a list of ictal clinical features that we considered atypical of mesial temporal lobe seizures (Table 1). We then examined the clinical history and the ictal video recordings of all seizures made during video EEG telemetry sessions in the patients from both groups and noted the occurrence of these atypical features. When a patient had more than one seizure recorded, we scored that patient for all atypical features occurring in any recorded seizure. To justify the individual inclusion of any sign in our analysis, we assessed also the strength of association (grouping) of individual signs.
The video telemetry recordings were scored independently on separate occasions by two different examiners (S.D.S., who was blinded to the radiographic findings and history, and P.B.). In general, there was agreement between the two observers as to the occurrence of the atypical signs. In less than 10% of seizures, the scoring of atypical features differed, and in these cases, the two examiners reviewed the tapes together and reached agreement about the clinical features.
The occurrence of an atypical feature was related to typical temporal lobe features. We defined an atypical feature as: “very early” if it occurred as the first sign or symptom (i.e., an atypical aura or bilateral tonic posturing with no aura first); “early” if it occurred after at least one mesial temporal sign or symptom (i.e., pedaling immediately after an abdominal rising sensation); “late” if it occurred after at least two mesial temporal features (i.e., temporal aura, loss of contact then pedaling of lower limbs); and “very late” if it occurred after signs indicative of a typical mesial temporal lobe seizure.
Surgical outcome was assessed using the Engel classification (Engel, 1987). The occurrence of atypical signs was compared in those with Engel class I, II, III, and IV outcomes. Surgical outcome was assessed at the last visit (a mean follow-up period of 4.7 years; range 1–9 years). Outcome was assessed in 61 of the 66 patients, and in the other five, follow-up was either too short or information incomplete for inclusion in the study.
Statistical comparison was made of: (1) the occurrence of individual atypical features in the two groups (severe HS versus mild HS) using the Pearson chi-square (χ2) test with the Yates's correction; (2) the number of patients in both groups with one, two to four, and five or more atypical features using the Mann Whitney U test; (3) the outcome of surgery in patients with or without atypical features using the Pearson χ2 test with the Yates's correction; and (4) the timing of the atypical features in relation to the occurrence of typical temporal lobe features (early or very early versus late or very late) and the postsurgery outcome (Pearson χ2 test with the Yates's correction).
Sixty-six adult patients, 36 females and 30 males, were included in the study. The average age of the whole sample was 36 years. The average age was 36 years (range 25–53) in the mild HS group and 37 years (range 20–57) in the severe HS group. The average age of the patients with atypical features was 37 years (range 20–57), and the average age of patients with no atypical features was 36 years (range 21–55). The average disease duration of the whole sample was 28 years. The average disease duration of the severe HS group was 30 years (range 13–53), and the average disease duration of the mild HS group was 27 years (range 7–44). Moreover, the average disease duration of the patients with atypical feature was 29 years (range 12–53) compared to 27 years (range 7–44) for the patients with no atypical features. Using the Mann Whitney U test, none of these differences turned out to be statistically significant (p value ranging between 0.36 and 0.62).
A history of febrile seizures (FS) was present in 20 out of 39 (51%) patients with severe HS compared to 16 out of 27 (59%) patients with mild HS. FS was reported in 17 out of 35 (48%) patients with atypical features and in 15 out 31 (48%) of the patients without atypical features. Neither of the latter two differences is significant statistically (p figures respectively 0.69 and 0.655, Pearson χ2 with Yate's correction).
We analyzed 188 seizures on video telemetry in these 66 patients. In these 188 seizures, 101 atypical features (87 signs and 14 symptoms) were observed (54% of seizures). Eighty-six of the 101 atypical features (75 signs out of 87 and 11 symptoms out of 14) were detected in patients with severe HS, and 15 (12 out of 87 signs and 3 out of 14 symptoms) were detected in the patients with mild HS (Tables 2 and 3). To justify considering signs and symptoms individually, we also examined the correlation between signs and symptoms. The correlation was weak for all atypical symptoms or signs, and individual features occurred independently.
Table 2. Frequency of atypical signs in the two patient groups (those with severe HS and those with mild HS)
|Atypical ictal signs|
| Asymmetric bilateral tonic posturing||11 (28%)||2 (8%)|
| Speech arrest or loud vocalization||11 (28%)||2 (8%)|
| Ictal facial grimacing||11 (28%)||1 (4%)|
| Complex partial seizures with prominent or agitated motor behaviors (i.e., pedaling, cycling, boxing, rocking movements)||15 (38%)|| 4 (15%)|
| Prominent head and eye tonic deviation|| 9 (23%)||2 (8%)|
| Clonic or tonic–clonic contralateral movements|| 7 (18%)||1 (4%)|
| Behavior with strong emotional or aggressive quality (i.e., crying, laughing)|| 7 (18%)||0 |
| Sudden onset and offset of ictal motor features|| 4 (13%)||0 |
|Atypical ictal symptoms|
| Aura with cephalic, body or throat sensation|| 7 (18%)|| 3 (11%)|
| Elementary sensory disturbance (i.e., tingling, numbness)|| 3 (8%) ||0 |
| Multimodal hallucinations|| 1 (3%) ||0 |
|Total (symptoms and signs)||86 ||15 |
Table 3. Number of abnormal signs/symptoms in individual patients
|No. of patients with one atypical sign/symptom|| 6||4|
|No. of patients with two to four atypical sign/symptoms||18||3|
|No. of patients with five or more atypical sign/symptoms|| 4||0|
Atypical ictal signs (motor features observed from review of the telemetry data) and/or symptoms (determined from chart reviews) were found in 35 out of 66 patients overall. Twenty-eight of 39 (72%) patients in the severe HS group had one or more atypical features compared to 7 of 27 (26%) of the mild HS group. This difference is statistically significant (χ2 test, p = 0.0013). Atypical ictal symptoms (elicited from the clinical notes) were found in 14 of 66 patients overall; 11 of 39 (28%) patients in the severe HS group had one or more atypical features compared to 3 of 27 (11%) patients in the mild HS group.
The mean number of atypical clinical features was found to be 2.2 in the severe HS group and 0.62 in the mild HS group. The frequency of the atypical features is shown in Table 2. This difference was statistically significant (Mann Whitney U Test, p < 0.001).
The outcome of temporal lobectomy was compared to the occurrence of atypical signs (Table 4). Of the 61 patients, 53 (87%) had an Engel class I outcome, 7 had a class 2 outcome (11%), and one had a class 3 outcome. Outcome was not statistically different in individuals showing any particular atypical features or in the number of individuals with atypical features (Table 4). However, both behavior with a strong emotional or aggressive quality (i.e., crying or laughing), clonic or tonic–clonic contralateral movements, and also ictal grimacing occurred almost exclusively in the Engel class I group. Given the high proportion of patients in the Engel class I outcome category, the power of this study to detect differences was low.
Table 4. Surgical outcome and its relationship to atypical features
|No. of patients with one atypical sign/symptom||9||1|
|No. of patients with two to four atypical sign/symptoms||18 ||3|
|No. of patients with five or more atypical sign/symptoms||2||2|
|Atypical ictal signs|| |
|Asymmetric bilateral tonic posturing||8||3|
|Speech arrest or loud vocalization||8||3|
|Ictal facial grimacing||9||1|
|Complex partial seizures with prominent or agitated motor behaviors (i.e., pedaling, cycling, boxing, rocking movements)||15 ||4|
|Prominent head and eye tonic deviation||7||3|
|Clonic or tonic–clonic contralateral movements||6||1|
|Behavior with a strong emotional or aggressive quality (i.e., crying, laughing)||7||0|
|Sudden onset or offset of ictal motor features||4||1|
|Aura with cephalic, body, or throat sensation||6||2|
|Elementary sensory disturbance (i.e., tingling, numbness)||3||0|
Of the eight patients whose surgical outcome was Engel class II or worse, two had typical temporal lobe features, five had mainly temporal lobe features (especially at the onset followed by a variable degree of extratemporal features), and one patient had seizures with exclusively extratemporal lobe semiology. Of the 53 patients with Engel class I outcome, four had seizures with predominantly extratemporal lobe features (one of these experienced a typical temporal lobe aura), while three had no temporal features at all. There is no significant statistical difference between the two outcome groups for the number lacking typical temporal lobe features (p = 0.4663).
Since the timing of an atypical feature may correlate with its relevance, we considered the temporal sequence of the atypical features (see the Methods section). Results are listed in Table 5. The majority of patients (21 of 35) had atypical features early (but not as the first symptom or sign) or late in their seizure evolution. Eight patients experienced an atypical feature as the first sign or symptom (seven of these had an atypical aura). Three of 16 patients with very early or early atypical features had an Engel surgical outcome of class II or worse compared to 3 of 19 of the patients having atypical features in the late or very late stages of the seizure. This difference is not statistically significant (p = 0.8169).
Table 5. Temporal distribution of atypical features in the context of the whole seizures
|8 (2)||8 (1)||13 (1)||6 (2)|
The core symptoms and signs of partial seizures arising in mesial temporal lobe structures are well described and are different from the core symptoms and signs from partial seizures arising in other brain regions. However, it is a common clinical observation that many patients with mesial temporal lobe epilepsy have some ictal features that are more characteristic of extratemporal lobe epilepsy. This phenomenon has not been extensively studied, and not much is known about the frequency, mechanisms, or anatomicoclinical correlations of atypical features in patients with temporal lobe epilepsy. Ching-wei et al. (2004) for instance noted marked lower limb motor features in 38% of the 123 temporal lobe seizures recorded in 38 patients who were seizure-free after temporal lobectomy. Carreno and colleagues (2005), however, observed complex motor behaviors in only 12 of 502 patients with temporal lobe epilepsy—of whom seven were operated upon and all seven became seizure-free. It seems likely that the occurrence of atypical features is due to the propagation of activity spreading beyond the temporal lobe borders (Salanova et al., 1995; Ching-wei et al., 2004) or the presence of a wide neuronal network underpinning the (so-called) mesial temporal lobe seizures involving extratemporal locations (Lieb et al., 1991; Bartolomei et al., 2002).
Two hypotheses are examined in this paper. The first is that seizures associated with severe HS were more likely to have extratemporal clinical features than are those associated with mild HS. To examine this, we used, as a population frame, those patients who had had a standard anterior temporal lobe resection—as it is in such patients that the presumption of seizure onset in the sclerotic temporal lobe is most firmly based. We compared, in this group, those patients who had severe hippocampal volume loss (ratio of <60%) with those who had mild volume loss (ratio of >75%) as two polar groups. Our findings confirmed the hypothesis. There was a striking difference in the occurrence of atypical features, which occurred in 72% of the patients with severe hippocampal volume loss compared with 26% of those with mild hippocampal volume loss. Furthermore, the patients with severe hippocampal volume loss had more atypical features compared to those with mild hippocampal volume loss. These figures of course might be influenced by potential selection bias (as only operated cases were included), but are nevertheless important, as these sorts of cases are exactly those undergoing surgical evaluation. It is also important to note that in most cases, the atypical features were not prominent—only 15 seizures (in five patients) were absolutely typical of frontal lobe attacks in the view of the senior author (S.D.S.). The more frequent occurrence of extratemporal features in those with more severe HS suggests the involvement of different networks, either because the severely sclerosed hippocampus plays a different role or because the degree of sclerosis is a marker of abnormalities beyond the hippocampus (Moran et al., 2001; Bonilha et al., 2006). Importantly, we found that the degree of volume loss was not significantly related to other factors, such as the duration of the disease, age of onset, or history of febrile seizures. We did not measure volumes of other extrahippocampal structures, although significant atrophy was not reported on visual inspection by an experienced neuroradiologist (J.M.S.). It is possible that the presence of extrahippocampal atrophy might correlate with atypical features, although a close structural (atrophy) and functional (network/spread) correlation would perhaps not be expected.
Another interesting observation from these data is that the atypical features were those usually associated with the prefrontal cortex, most probably reflecting the extensive connections between the mesial temporal lobe, limbic system, and the prefrontal cortex (Carmichael & Price, 1995). The commonest atypical features were the complex partial seizure with pedaling, cycling, boxing, etc., occurring 19 times overall (15 times in the severe HS group and 4 in mild HS group) and followed by the sudden and violent vocalization or speech arrest feature, the asymmetric tonic posturing feature, and the ictal facial grimacing feature (see Table 2). Other symptoms such as version and atypical tonic posturing, typically associated with supplementary motor cortex, were also commonly observed.
The second hypothesis examined is that the outcome of temporal lobectomy (standard anterior temporal lobe resection) in patients in whom preoperative seizures had extratemporal features is less successful. If this were the case, it might reflect the larger neuronal network with an epileptogenic zone that extends into extratemporal locations or even origin of the seizures outside the temporal lobe (Ryvlin & Kahane, 2005). A (simplistic) view might be that mesial temporal lobe resection in this situation would be expected to be less successful. We did not find this. In our study, the outcome was not strikingly different in those with or without extratemporal features, and furthermore, no feature was particularly associated with a good or bad outcome. Similarly, in the study of Ching-wei et al. (2004), the seven operated cases all did well, and in the study of Carreno and coworkers (2005), all the 38 cases were seizure-free (an inclusion criterion). At our center, we carry out anterior lobe resection and resection of the mesial temporal structures; this is important, because both parahippocampal structures and temporal pole can be affected in patients with HS (Moran et al., 2001; Coste et al., 2002; Bernasconi et al., 2003). Our good outcome may therefore relate to our relatively large extent of resection (Bonhila et al., 2007). It remains to be seen whether more limited mesial temporal lobe resections in patients with extratemporal features gives a similarly good outcome. Further, as most patients achieved seizure freedom in our study, the power of the study to detect a difference between groups was low. Having said this, it is clear that surgery in patients with atypical clinical features, providing other investigatory findings are concordant, is usually successful. The clinical point here is that the occurrence of atypical features is not in itself a contraindication to surgical therapy.
We would like to thank Karl Ng for his help in initiating this study.
Conflict of interest: We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. The authors declare no conflicts of interest.