- Top of page
Summary: Purpose: We investigated the response to antiepileptic drug (AED) therapy in patients with localisation-related epilepsy associated with different underlying causes.
Methods: Five hundred and fifty adolescent and adult patients who had partial epilepsy treated with AEDs and who had undergone magnetic resonance imaging of brain were followed up prospectively from 1984 at a single centre. More than 70% were newly diagnosed. None had had epilepsy surgery.
Results: Three hundred and twelve (57%) patients had been seizure free at their last clinic visit for at least a year. Patients with mesial temporal sclerosis (MTS; n = 73, 42% seizure free) were less likely to be controlled (p < 0.01) than were those with arteriovenous malformation (AVM; n = 14, 78%), cerebral infarction (n = 46, 67%), primary tumour (n = 35, 63%), cortical gliosis (n = 81, 57%), cerebral atrophy (n = 49, 55%), and cortical dysplasia (CD; n = 63, 54%). Among the seizure-free patients, those with MTS were more likely to require more than one AED compared with those with other aetiologies (48 vs. 35%; p < 0.05). There was no difference in outcome between patients with symptomatic and cryptogenic epilepsy (n = 361, 58% vs. n = 189, 56% seizure free, respectively). Patients with MTS, CD, and cryptogenic epilepsy were more likely (p = 0.02) to have a family history of epilepsy than were the other groups. MTS patients also had a higher incidence of febrile convulsions (p < 0.001).
Conclusions: The majority of patients with focal-onset epilepsy became seizure free on AED treatment. MTS-related seizures had the worst prognosis. Although many patients with this pathology may benefit from epilepsy surgery, a considerable number will be controlled with AED therapy.
Epilepsy is the most common serious neurologic disorder. It is estimated to affect up to 50 million people worldwide (1). Although the prognosis is generally good, more than 35% of patients continue to have seizures despite appropriate deployment of antiepileptic drug (AED) therapy (2). In the last decade, considerable progress has been made to classifying seizure types (3) and epilepsy syndromes (4). Advances in brain imaging have allowed better identification of the structural abnormalities underlying localisation-related epilepsy (5,6). These commonly include mesial temporal sclerosis (MTS), cortical dysplasia (CD), cortical gliosis, atrophy and infarction, primary neoplasm, and arteriovenous malformation (AVM).
Localisation-related epilepsy is less easy to control with AEDs than are the primary generalised epilepsies (2). Many affected patients develop refractory partial-onset seizures, which may be abolished or better controlled after resective surgery (7). Patients with MTS-induced seizures are particularly likely to have hippocampal excision (8–10) because of a perceived resistance to AED treatment (9,11–13). CDs are also widely thought to be associated with refractory seizures, which also may be amenable to surgery (14). Published data are thus often concerned with surgical outcome in patients with partial-onset seizures, with only a handful of studies examining the response to AED therapy. This is despite the global introduction in the last decade of nine novel AEDs for this indication (15). This study examined prospectively the outcome in pharmacologically treated patients with partial epilepsy secondary to a range of structural abnormalities.
- Top of page
The subjects consisted of adolescents and adults with localisation-related epilepsy referred to the epilepsy clinic at the Western Infirmary in Glasgow, Scotland, from January 1, 1984, until December 31, 1997. At the first visit, demographic and clinical information was obtained from each patient and any witness to the seizures using a standardised structured questionnaire. Information collected included a detailed description of seizures, previous AED treatment, and risk factors for epilepsy such as febrile convulsions, head trauma, and family history of epilepsy in first- and second-degree relatives. Patients in whom the clinical history or electroencephalographic (EEG) results suggested localisation-related epilepsy underwent routine magnetic resonance imaging (MRI) of the brain in the 7 years before analysis.
A standardised brain MRI protocol was used with a 1.0-T Siemens Magnetom Scanner with a head coil. T2-weighted turbo-spin-echo sequences were obtained to cover the whole brain in the axial plane [repetition time (TR) = 4,000 ms; echo time (TE) = 120 ms; flip angle = 180°; section thickness = 5 mm; intersection gap = 0.2 mm; field of view (FOV) = 230 mm; echo train = 23]. T1-weighted MPRAGE sequences were acquired in the coronal plane (TR = 11.4 ms; TE = 4.4 ms; flip angle = 12°; slab thickness = 250 mm; effective thickness = 1.98 mm; FOV = 250 mm). Further 4-mm-thick high-resolution T2-weighted coronal images were acquired with TR = 4,465 ms; TE = 120 ms; flip angle = 180°; FOV = 230 mm; matrix size 300 × 512. The coronal planes were perpendicular to the longitudinal axis of the hippocampal body.
MRIs were assessed qualitatively (11,16,17), and all films were reviewed blind by the same neuroradiologist to ensure the veracity of the original report. Abnormalities were classified as CD (abnormalities of gyration, heterotopia, tuberous sclerosis, focal cortical dysplasia, microdysgenesis, dysembryoplastic neuroepithelial tumour, megalencephaly / hemimegalencephaly) (18); MTS; cortical gliosis (posttraumatic brain injury); cerebral infarction; tumours (glioma, meningioma); AVM; and cerebral atrophy. As only five patients had dual pathology (MTS and another structural lesion), these were excluded from the study, as were those with surgically resectable tumours.
Patients were reviewed initially every 4–6 weeks for 6 months and subsequently at least every 4 months for a minimum of 2 years. Compliance was monitored with the aid of on-site measurement of plasma drug concentrations at the clinic (19). Demographic and clinical data, results of investigations, and response to treatment for each patient were entered systematically into an ongoing prospective database (20).
Seizure types (3) and epilepsy syndromes (4) were classified according to the guidelines of the International League Against Epilepsy. The epilepsies were broadly classified as generalised or localisation-related (partial), each of which was further subclassified as idiopathic, symptomatic, or cryptogenic, according to the putative aetiology, taking into account the age of the patient, type of seizures, presence or absence of a family history of epilepsy, and other clinical characteristics. Only patients with localisation-related epilepsy, in whom seizure semiology or investigation findings suggested a focal origin, were included in this study. Symptomatic epilepsies were considered to be the consequence of a structural abnormality identified on brain imaging. Cryptogenic epilepsies were presumed to be of partial onset, based on the clinical description of the seizures and surface EEG, but with no structural abnormality identified on brain imaging.
Patients were divided into two groups according to whether they were seizure free or uncontrolled. Seizure freedom was defined as the absence of any type of seizures or auras for a minimum of 1 continuous year. Seizure control was assessed at the time of the patient's last clinic visit. The χ2 test was used for comparisons of categoric data. All statistical tests were two-tailed. Calculations were computed using Minitab for Windows (Release 11.21) software. Outcome was compared between patients with MTS and those with other pathologies combined as a group.
- Top of page
Five hundred fifty patients (50% male and 50% female patients) with localisation-related epilepsy entered the study. Of these, 385 (70%) were newly diagnosed, whereas the remaining 165 (30%) were taking AED treatment at time of referral. The median age at onset of epilepsy was 21 years (range, <1–92 years). The median age at referral was 30 years (range, 12–93 years). The median follow-up period was 5 years (range, 2–15 years). Structural abnormalities were identified on brain MRI in 361 (66%) patients, who were classified as having symptomatic epilepsy. Eighty-one (15%) patients had cortical gliosis, 73 (13%) MTS, 63 (12%) CD, 49 (9%) cerebral atrophy, 46 (8%) cerebral infarction, 35 (6%) primary tumour, and 14 (3%) AVM. MRI was normal in the remaining 189 patients (34%), who were classified as having cryptogenic epilepsy. No patient was diagnosed as having idiopathic localisation-related epilepsy in our cohort of adolescent and adult patients.
Overall, 312 (57%) patients were seizure free at the time of analysis. There was no significant difference in seizure free rate between the newly diagnosed patients (59%) and those already receiving treatment at the time of referral (50%). There was no difference in outcome between patients with symptomatic or cryptogenic epilepsy (56 vs. 58% seizure free, respectively). Patients with MTS (n = 73; 42% seizure-free) were less likely to be seizure free (Fig. 1; p < 0.01; relative risk, 1.4; 95% confidence intervals, 1.1–1.8), than were those with AVM (n = 14, 78%), cerebral infarction (n = 46, 67%), primary neoplasm (n = 35, 63%), cortical gliosis (n = 81, 57%), cerebral atrophy (n = 49, 55%), and CD (n = 63, 54%). There was no statistical difference in outcome among patients with other lesions. Of the 63 patients with CD, 34 (54%) were seizure free. Thirty-six categories of dysplastic lesion were identified on brain MRI, the commonest being right temporal heterotopia (n = 7) and bilateral periventricular tubers (n = 6). No specific dysplastic lesion was associated with a particularly poor prognosis.
Figure 1. Seizure control in different patient groups. MTS, mesial temporal sclerosis; AVM, arteriovenous malformation.
Download figure to PowerPoint
Of the 312 seizure-free patients, 187 (60%) patients were receiving treatment with a single AED (Table 1). Eighty-three (27%) patients required two AEDs, whereas 15 (4.7%) were controlled with three drugs, and one took four. Thirty-four patients elected not to take an AED, 26 having never been treated, and eight having discontinued therapy because of intolerable side effects. Of these 34 individuals, 26 experienced only sporadic seizures, with none in the previous year, whereas the remaining eight had more frequent seizures. Among the seizure free patients, 48% of patients with MTS required more than one AED (p < 0.05; relative risk, 1.6; 95% confidence intervals, 1.1–2.4) compared with 35% of those with other aetiologies (20% with CD, 23% with primary tumour, 27% with AVM, 29% with infarction, 30% with gliosis, 44% with atrophy, and 31% with cryptogenic epilepsy). Patients with uncontrolled epilepsy also took a range of AEDs, varying from none to four (Table 2).
Table 1. Seizure-free patients on no treatment and on different antiepileptic drug (AED) regimens
| ||Patients n (%)||No AED n (%)||1 AED n (%)||2 AEDs n (%)||3 AEDs n (%)||4 AEDs n (%)|
|MTS||31 (100)||7 (23)||9 (29)||12 (39)||2 (6)||1 (3)|
|CD||34 (100)||5 (15)||22 (65)||6 (17)||1 (3)||0 (0)|
|Atrophy||27 (100)||1 (4)||14 (52)||11 (40)||1 (4)||0 (0)|
|Gliosis||46 (100)||3 (7)||29 (63)||12 (26)||2 (4)||0 (0)|
|Tumour||22 (100)||1 (5)||16 (72)||4 (18)||1 (5)||0 (0)|
|Infarction||31 (100)||1 (3)||21 (68)||7 (22)||2 (7)||0 (0)|
|AVM||11 (100)||1 (9)||7 (64)||2 (18)||1 (9)||0 (0)|
|Cryptogenic||110 (100)||7 (6)||69 (63)||29 (27)||5 (4)||0 (0)|
|Total||312 (100)||26 (8)||187 (60)||83 (27)||15 (4.7)||1 (0.3)|
Table 2. Patients uncontrolled with no treatment and with different antiepileptic drug (AED) regimens
| ||Patients n (%)||No AED n (%)||1 AED n (%)||2 AEDs n (%)||3 AEDs n (%)||4 AEDs n (%)|
|MTS||42 (100)||1 (2)||7 (17)||24 (57)||10 (24)||0 (0)|
|CD||29 (100)||2 (7)||10 (34)||13 (45)||3 (10)||1 (4)|
|Atrophy||22 (100)||1 (4)||9 (41)||8 (37)||3 (14)||1 (4)|
|Gliosis||35 (100)||1 (3)||15 (43)||11 (31)||8 (23)||0 (0)|
|Tumour||13 (100)||0 (0)||2 (15)||7 (54)||4 (31)||0 (0)|
|Infarction||15 (100)||0 (0)||7 (47)||6 (40)||2 (13)||0 (0)|
|AVM||3 (100)||0 (0)||1 (33)||2 (67)||0 (0)||0 (0)|
|Cryptogenic||79 (100)||3 (4)||29 (36)||41 (52)||6 (8)||0 (0)|
|Total||238 (100)||8 (3)||80 (34)||112 (47)||36 (15)||2 (1)|
Family history of epilepsy in first- and second-degree relatives was more common (Table 3; p = 0.02) in patients with MTS, CD, and cryptogenic epilepsy (all 25%) and AVMs (28%) than in those with atrophy (20%), infarction (15%), gliosis (9%), and neoplasia (6%). The presence of a family history seemed to convey a better prognosis in the MTS patients (present: n = 18, 61% seizure free; absent: n = 55, 36% seizure free; p = 0.065). MTS patients were more likely to have had febrile convulsions (Table 3) than were those in the other groups (p < 0.001; relative risk, 4.1; 95% confidence intervals, 2.7–6.2), although this did not appear to affect outcome (present: n = 27; 41% seizure free; absent: n = 46, 43% seizure free).
Table 3. Family history of epilepsy and febrile convulsions in different patient groups
| ||MTS||CD||Atrophy||Gliosis||Tumour||Infarction||AVM||Normal MRI|
|Number of patients||73||63||49||81||35||46||14||189|
- Top of page
More than half of our patients with partial epilepsy became seizure free with AED treatment. The majority were taking monotherapy. MTS patients were significantly more likely to be refractory and to require more than one AED to attain seizure freedom. Within this pathologic grouping, however, there was substantial variation in seizure control, ranging from seizure freedom taking no AEDs to refractory with four. This phenomenon was observed in patients with other diagnoses. Complete seizure control was achieved in the majority of patients with CD. No difference in outcome was found between patients with symptomatic or cryptogenic epilepsy. The latter had similar rates to patients with CD for seizure freedom and epilepsy risk factors. A family history of epilepsy was more common in patients with MTS, CD, cryptogenic epilepsy, and AVM. This characteristic seemed to convey a better prognosis for patients with MTS. A history of febrile convulsions was statistically more prevalent in MTS patients compared with those with other pathologies.
Consistent with other reports (9,11–13), patients with MTS had the poorest prognosis. There are few data, however, with which these outcomes can be compared. Semah et al. (12) studied seizure frequency >1 year in a cohort of 1,369 outpatients with localisation-related epilepsy. As in our study, patients with infarcts, vascular malformations, and tumours had the best prognosis, with 54, 50, and 46% being seizure free, respectively. Their outcomes for MTS patients were poorer, with only 11% being fully controlled. Only 8% of the cohort were newly diagnosed, however, and the remaining pharmacoresistant patients contributed to the low seizure-freedom rates. Van Paesschen et al. (13) examined for 1 year 63 outpatients with seizures who underwent brain MRI. Of these, all six with MTS remained refractory despite AED treatment, unlike 57 with other pathology, 30 of whom became seizure free. Kim et al. (11) reported that 25% of 104 outpatients taking AED treatment for MTS remained seizure free during 2 years' follow-up. Again, the majority were previously taking AEDs at time of referral, with only 15% being newly diagnosed. Briellmann et al. (9) also published a report on 47 patients with MTS being assessed for surgery. Not surprisingly, none was seizure free.
More than half of our patients with CD attained complete seizure control. These patients fared better than those studied by Semah et al. (12), who observed that only 24% of their patients with this diagnosis were seizure free. Palmini et al. (14) examined 30 patients with CD-related refractory seizures. All but five had never experienced remission of their seizures. This study may have selected a particularly pharmacoresistant population, however, because these patients were being considered for epilepsy surgery. The variability in outcome within our group of CD patients may reflect the wide range of lesions that comprise the condition (21–23).
Although our MTS and CD patients had a more promising outlook than those previously studied, there was substantial heterogeneity among this group, with some patients in remission without treatment, whereas others continued to have seizures despite taking up to four AEDs. Murakami et al. (24) observed this variability in children with MTS. The phenomenon may be a reflection of the natural history of these heterogeneous conditions, although this is difficult to explore, given their low incidence and the fact that the clinical course may be altered by AED therapy and/or surgery. The reporting at autopsy of MTS among nonepileptic subjects supports this hypothesis (25–27).
A family history of seizures was present in 25% patients with MTS. Interestingly, this seemed to convey a better prognosis, suggesting perhaps a hereditary component to the disease process in some patients (28). Such an association has been noted in patients who underwent temporal resection, although they had a variety of pathologies (29). Febrile convulsions occurred in 37% of patients with MTS, a higher figure than in all other groups. This is a well-recognised relationship (28). A history of febrile convulsions did not affect outcome in our study. In previous studies, this has been associated with poorer response to AED treatment (11), but better outcome after temporal resection (29). It has been postulated that a preexisting hippocampal lesion may predispose some individuals to febrile convulsions, MTS, and a tendency to refractory epilepsy, rather than febrile convulsions being the causative factor (28,30,31).
Patients with CD and AVM had a higher prevalence of family history of seizures than did those with infarcts, gliosis, and neoplasia. CD may be the result of familial disorders such as tuberous sclerosis or neurocutaneous syndromes (18). MTS with duplication/dispersion of the dentate fascia may be a form of CD (32). The association of family history in patients with AVMs may be explained by a previous MRI finding that up to 75% sporadic cases of cavernous angiomas are familial (33).
Characteristics of patients with normal brain imaging were similar to those with CD, in terms of percentages of seizure free patients (58 and 54%, respectively), those with a family history of epilepsy (25% each), and those with a history of febrile convulsions (12 and 11%, respectively). This lends credence to the suggestion that some patients with cryptogenic epilepsy may have cortical microdysplasia. Microdysgenesis can increasingly be identified with sophisticated neuroimaging techniques (34).
In conclusion, the outcome in patients pharmacologically treated for symptomatic or cryptogenic epilepsy was highly variable. Whilst MTS was associated with the poorest prognosis, many of our patients with the diagnosis were seizure free. Encouraging results were also obtained for patients with CD. These data provide a balance to the perception that epilepsy due to MTS or CD is likely to be refractory to AED treatment. This is often a self-fulfilling prophecy, because such patients tend to be assessed in centres specialising in epilepsy surgery. Substantial prospective studies are required to explore the clinical course of pharmacologically treated patients with newly diagnosed partial epilepsy.