The Irish epilepsy surgery experience: Long-term follow-up
Article Outline
- Abstract
- 1. Introduction
- 2. Methods
- 3. Results
- 4. Discussion
- Conflicts of interest
- Funding
- References
- Copyright
Abstract
Aim
To assess the long-term seizure outcome of Irish patients who underwent resective surgery for refractory epilepsy since 1975. We also wished to determine the impact of pathology and surgical technique (with particular reference to neocorticectomy) on seizure outcome.
Methods
A retrospective review of medical notes, radiological and histopathological records, was undertaken between 1975 and 2005. Missing data was supplemented by telephone calls to patients. One hundred and ninety-nine patients suited the criteria for inclusion and had at least 1-year follow-up (1–24 years, mean 7.0 years). Engel's criteria were used to classify seizure outcome at 1, 2, 5, 10, 15 and >15 years follow-up.
Results
The percentage of patients seizure free at 2, 5, 10, 15 and >15 years were, 56.6%, 41.4%, 44%, 25% and 31.3%, respectively. Of patients with a pathologically confirmed diagnosis of mesial temporal sclerosis, 55.6% were seizure free at 10 years. Equivalent figures for tumour were 62.5%, for cortical dysplasia, 34.8%, for those without any demonstrable pathologic abnormality, 50%, for dual pathology, 50% and for all others, 33.3%. Of those with 10 years or greater follow-up only 20% of neocorticectomy patients were in Engel class 1, compared with an average of 58.5% for the other surgical techniques.
Conclusion
Seizure freedom rates for Irish Patients were comparable to other large retrospective studies. Patients who underwent selective procedures tended to do better than those undergoing lobar resections, in keeping with international trends. The surgical technique unique to the Irish cohort, temporal necocorticectomy, had the worst long-term outcome.
Keywords: Irish epilepsy surgery series, Long-term outcome, Pathology outcome, Neocorticectomy
1. Introduction
This is the first comprehensive audit of the Irish epilepsy surgery series since the first temporal lobectomy was performed in Dublin in 1975. The National Centre for Epilepsy Surgery was officially established in Beaumont Hospital, Dublin in 1994 and is the only centre in Ireland that offers surgical treatment of epilepsy.
The effectiveness of surgery over continued pharmacotherapy for intractable patients has been established in a recent randomized controlled trial1 which highlighted the short-term efficacy of resective surgery. Long-term outcome studies, however, are valuable for both pre- and post-surgical patients, to allow a fully informed choice for the former and a realistic expectation of the future for the latter.
While the evolution of epilepsy surgery techniques in Ireland reflected international trends, the origins of the Irish series are quite unique. Neocorticectomy was a popular surgical technique when epilepsy surgery was first offered in Dublin (1975–1992) and this centre provides the only substantial modern series of this technique. A 1992 study2 reported outcome results for neocorticectomy patients, and this now allows us to compare this unique surgical technique with the more typical methods of surgery adopted for epilepsy which succeeded it.
The focus of this study is the long-term post-operative seizure outcome of patients who underwent resective surgery for epilepsy. The pathological diagnoses and surgical techniques used (with particular reference to neocorticectomy) are also examined to determine their impact on outcome.
2. Methods
2.1. Sample
The total number of patients who underwent surgery for epilepsy in Ireland since the first operation in 1975 is unknown as initially formal records were not maintained. Apart from poor initial record keeping, data collection was hampered by other issues such as the relocation of the neurosurgical programme in Dublin to the new Beaumont Hospital in the late1980s, and the lack of national recognition of the epilepsy programme until the 1990s. Patient names were gathered from various sources including previous epilepsy research studies at our institution and old operating room log books. Three hundred and twenty-nine patients who had undergone surgery for epilepsy were identified. Patient criteria for inclusion in this study were as follows: (1) patient had a pre-operative diagnosis of refractory epilepsy, (2) had resective surgery with the aim of eliminating seizures or reducing their frequency or severity, (3) surgery was resective not functional (e.g. functional hemispherectomy or insertion of a vagal nerve stimulator) and (4) at least 1-year follow-up information was available.
Of the 329 patients identified hospital records were available for 272 patients. A further 44 patients were excluded according to the above criteria: 19 patients had acute symptomatic seizures in the immediate peri-operative period and the aim of their surgery was removal of a tumour causing mass effect; 15 patients did not have medical intractability established pre-operatively, and 10 patients had functional surgical procedures leaving 228 patients for analysis.
2.2. Design and data collection
A retrospective chart review was undertaken. Of the 228 patients, 150 had incomplete chart follow-up information. The Medical Research Ethics Committee at Beaumont Hospital gave approval to contact patients by telephone when follow-up information was incomplete. A letter was first sent to the patient's general practitioner informing him/her of the study and asking if there was any reason that it would be inappropriate to attempt to contact the patient (e.g. if the patient had died). Patients were then sent an information leaflet explaining the aims and objectives of the study, and requesting that a consent form be returned stating whether or not they wished to be contacted by telephone. Of the 150 patients with incomplete information, only 50 were successfully contacted. The remaining 100 patients were still included in the study but the number of years of follow-up was less than it would have been had recent contact been possible. Of the 228 patients who met the inclusion criteria for this study, there was no post-operative follow-up for 28 patients and 1 patient died peri-operatively, leaving 199 patients with follow-up of 1 year or more (1–24 years of follow-up, mean 7.0 years) (Fig. 1).
Fig. 1.
Epilepsy surgery patients.
2.3. Follow-up outcome information
Outcome information, with regard to seizures, was classified using Engel's criteria3 at 1, 2, 5, 10, 15 and >15 years. Class 1 included patients who were completely seizure free, had auras only or seizures secondary to withdrawal of anti-epileptic drugs. Class 2 patients had rare disabling or nocturnal seizures only, class 3 patients had a worthwhile seizure reduction, and class 4 included patients who had an unsatisfactory reduction in seizure frequency, an unchanged seizure profile, or exacerbation of seizures after surgery.
2.4. Seizure free definition
At each of the post-surgical anniversaries the patient's seizure status for the year preceding the anniversary was used to assign them to the appropriate Engel outcome class. For example, a patient who was followed up for 15 years and remained in Engel class 1 for each assessment date may not have been totally free of seizures for the whole duration of follow-up. The patient may have had seizures during years 2–4 after surgery but entered remission the year before the year five anniversary, and was therefore classed as seizure free at year five.
2.5. Pathology
Pathological diagnoses were accepted only if the original pathology report or a copy of it was available and signed by the neuropathologist. Patients were classified into one of the following seven categories (1) mesial temporal sclerosis—MTS (MTS described qualitatively on the basis of cell loss and gliosis), (2) cortical dysplasia—CD, (3) tumour, (4) vascular malformation, (5) no abnormal histology, (6) other, and (7) dual pathology.
2.6. Surgical technique
The operation note signed by the neurosurgeon, or a copy of it, was used to determine the surgical technique. Patients were classified into 1 of 5 categories: (1) anterior temporal lobectomy—ATL, (2) amygdalo-hippocampectomy—AH, (3) neocorticectomy, (4) lesionectomy, and (5) frontal lobe resection. The technique used for neocorticectomy has been previously described.4
3. Results
3.1. Overall outcome
One hundred and ninety-nine patients suited the criteria for inclusion in this study and had at least 1-year follow-up (1–24 years; mean 7.0 years). Table 1 displays the clinical characteristics of these patients.
Table 1. Clinical characteristics of patients.
| Characteristic | Number |
|---|---|
| Gender | |
 Female | 104 |
| Male | 95 |
| Age of epilepsy onset (years) | |
| Mean (SD) | 10.7 (8.9) |
| Range | 1 month–48 years |
| Duration of epilepsy (years) | |
| Mean (SD) | 15.8 (10.2) |
| Range | 6 months–43 years |
| Age at surgery (years) | |
| Mean | 26.3 (10.9) |
| Range | 1–61 years |
| Side of resection | |
| Left | 86 |
| Right | 113 |
The seizure outcome for all patients at 2, 5, 10, 15 and >15 post-operative years respectively is shown in Fig. 2.
Fig. 2.
Seizure outcome for all patients (n
=199).
3.2. Seizure freedom
Forty-three percent of patients in our study never relapsed, that is, they remained in Engel class one for the duration of their follow-up.
3.3. Late relapse
Thirty-four patients (17%) experienced a late relapse of seizures i.e. seizure recurrence following seizure freedom in the first one or two post-operative years. Twenty of these patients had follow-up of at least 1 year after relapse and only three (15%) regained their seizure free status and were classed as such at last follow-up (3–17 years, mean 9.25).
3.4. Long-term follow-up (15 years or more)
Thirty-two patients had follow-up for 15 years or more (16–24 years, mean 19.3 years) following surgery. Of these, 53.1% (n=17) had a favourable outcome (Engel's class 1 or 2) at last follow-up and 46.9% (n=15) an unfavourable outcome (Engel's class 3 or 4). Of the former group at last follow-up, 70.6% (n=12) had a favourable outcome at each follow-up anniversary and 29.4% (n=5) were seizure free for the entire duration of follow-up. Of the latter group at last follow-up, 60% (n=9) had been classed as unfavourable at each follow-up stage and 33.3% (n=5) were classed as late relapsers. Two patients with follow-up of 15 years or more exhibited the “running down phenomenon” (experienced seizures for the first 1–2 post-operative years and subsequently remitted).
3.5. Pathology
Of the 199 patients with at least 1-year follow-up, 194 had a neuropathological diagnosis that suited the criteria for inclusion. The most common pathology was MTS (n=73) followed by CD (including neuronal ectopia, heterotopia and neuronal clustering) (n=47). The remaining pathologies were divided as follows: tumour (n=23), vascular (n=5), no abnormal histology (n=13), other (n=15), and dual pathology (n=18). Of the 18 patients with dual pathology—12 had CD and MTS, 2 had astrocytoma and CD, 2 had ganglioglioma and CD, 1 had oligodendroglioma and CD and 1 had dysembryoplastic neuroepithelial tumour (DNET) and CD. Of the 15 patients with ‘other’ pathology classification—7 patients had Chaslin's gliosis insufficient for diagnosis of MTS, 2 had cerebral contusions, 2 had evidence of infarction, 1 had non-specific chronic inflammatory change, 1 had neurosarcoidosis, 1 had cortical hamartoma and1 had a porencephalic cyst. Of the 23 patients with tumours, 12 had astrocytomas, 4 had gangliogliomas, 4 had oligodendrogliomas, 2 had oligoastrocytomas and 1 had a DNET.
Fig. 3, Fig. 4 depict the seizure outcome for patients with different pathological diagnoses at post-operative years five and ten. Favourable seizure outcome included patients in Engel's class 1 or 2 and unfavourable, those in Engel's class 3 or 4. Vascular pathology is not included in this figure as this pathology group had follow-up for 1 year only.
Fig. 3.
Favourable seizure outcome (Engel classes 1 and 2) at 5 and 10 post-operative years
Fig. 4.
Unfavourable seizure outcome (Engel classes 3 and 4) at 5 and 10 years.
Pathology results were compared to determine if there was a significant difference between favourable and unfavourable outcome at each of the surgical anniversaries. MTS patients had a significantly better outcome over CD, Tumour and other pathology groups at 1, 2 and 5 years after surgery. There was a borderline significantly better outcome for MTS patients compared with patients in whom a pathologic abnormality was not demonstrated over the period of the 1–10-year anniversaries (p<0.06–0.04). Dual pathology had significantly better results over CD at 1 and 5 years, over vascular pathology at 1 year only, and over other pathology group at 2 years.
3.6. Surgical technique
Of the 199 patients with at least 1-year follow-up, the surgical technique suited the criteria for inclusion for 195 patients. ATL was the most common operation (n=74) followed by AH (transcortical approach) (n=62), neocorticectomy (n=34), lesionectomy (n=21) and frontal lobe resection (n=4). The latter two groups therefore including patients with extra-temporal lobe epilepsy. The seizure outcome for the various surgical techniques is shown in Fig. 5. Patients with MTS pathology who underwent AH surgical technique had significantly better outcomes than those who underwent ATL at the 2-, 5- and 10-year surgical anniversaries (p<0.006, p<0.003, p<0.006 respectively).
Fig. 5.
% Patients seizure free according to surgical technique at each surgical anniversary.
3.7. Neocorticectomy
The percent of neocorticectomy patients with a favourable outcome at one, two and five post-operative years was 53%, 51.7% and 42.3%, respectively. This increased to 60% at 10 years, 66.7% at 15 years and 77.7% at greater than 15 years. The most dramatic difference in outcome results for surgical techniques was between neocorticectomy and AH. AH had significantly better outcomes over neocorticectomy for follow-up periods 1–10 years (p<0.0005). While 91.2% of neocorticectomy patients experienced a relapse at some point during follow-up, only 35.5% of AH and 68.9% of ATL patients did so. The outcome differences between neocorticectomy and ATL were not so distinct. At the fifth post-operative anniversary significantly more patients from the ATL group were in Engel's class 1 compared to the neocorticectomy patients (p<0.02). The difference did not reach significance at the 1- (p>0.24), 2- (p=0.055) or 10-year anniversaries (p<0.08).
3.8. Mortality
Thirteen of the 199 patients (6.5%) died during the follow-up period. The cause of death is known in 5 patients. One patient died from a post-operative haematoma following epilepsy surgery, one from complications secondary to a seizure, another from a non-neurological cancer and two from suicide. The mean age at the time of death for the 13 patients was 35.5 years (range 18–60 years). Death occurred between 0 and 20 years after surgery (mean 8.3 years). Five of those who died had been seizure free for at least 2 years before their death.
4. Discussion
This analysis of the Irish epilepsy surgery series demonstrates that the seizure outcome for Irish patients is favourable and overall compares well to those from other countries. It also suggests that those experiencing late relapse tend to have a poor outcome with regards to achieving complete seizure control, as do those who underwent neocorticectomy.
Of patients with 5 and 10 years of follow-up, MTS patients were more likely than any other pathology group to have a favourable outcome (Engel class 1 or 2). Also there was very little change in the outcome of MTS patients between the 5th and 10th anniversaries. Conversely, patients in the pathology group ‘other’ were least likely to have a favourable outcome at the same anniversaries (Fig. 3, Fig. 4). CD and tumour pathologies were the only groups to have more patients with a favourable outcome at 10 years compared with 5 years. The most dramatic deterioration in outcome was for the dual pathology group, as 81.3% had a favourable outcome at 5 years but this dropped to 62.5% at 10 years.
The seizure free outcomes for our MTS patients were 75.3%, 72.1%, 59.5%, 55.6% at 1, 2, 5 and 10 years respectively. The outcome for MTS patients over the same time period in two other studies5, 6 were 67.7%, 62.2%, 54.2% and 47% and 88.8%, 88.8%, 91.3% and 85.7% respectively. The latter study also reported the 1-year seizure free outcome for MTS patients for a range of studies from 1995 to 2004 (range: 58–88%, mean 81.5%). Overall 43% of patients in our study remained seizure free throughout; this is consistent with results from other studies that range from 45.6 to 55%.7, 8, 9, 10
Patients without any demonstrable pathologic abnormality had similar outcome results to those with MTS. This is in contrast to other studies11, 12 that report that the absence of pathologic change predicts a poorer outcome compared with other pathologies. However this was disputed in other reports.13, 14
In terms of surgical technique, patients who underwent an AH had very good short- and long-term outcomes. Throughout the 1–10-year follow-up at least 77.4% of these patients were classed as seizure free, and had the highest number seizure free at the 10-year anniversary (84.6%). The significantly better outcome of our MTS patients who underwent AH compared to ATL (41% seizure free at 10 years) is in contrast to other recent studies which reported the opposite15 and no significant difference in outcome.16, 17 Patients who underwent AH did not experience a deterioration in seizure outcome with longer follow-up whereas those who underwent ATL were less likely to be seizure free at 10 years follow-up than at 1 year. There is no identifiable reason for this discrepancy and our outcome results for ATL patients are almost identical to those from another large scale study which also had follow-up of patients who underwent ATL for 15 years or more.5 The decision for patients to undergo AH or ATL was decided on a case by case basis, following the first AH at our institution which was performed in 1992. All surgical cases were discussed at the weekly multi-disciplinary epilepsy meeting. If a patient had clear clinical and EEG evidence of mesial temporal lobe seizure onset, and sclerosis on MRI imaging, they would be offered the selective technique.
With regards to achieving seizure freedom neocorticectomy patients did not do as well as those who underwent other surgical techniques. The vast majority (91.2%) of neocorticectomy patients experienced a relapse (perhaps due to the persistence of the mesial temporal structures). An improvement was seen at follow-up greater than 15 years (16–20 years, mean 17.8 years) with 33.3% of patients seizure free. This may have been due to the introduction of newer anti-epileptic drugs in these patients. Neocorticectomy was performed in Dublin when resective surgery for epilepsy was first introduced. The outcome of 50 Irish neocorticectomy patients had previously been published in 1992.2 This study had followed up patients from 3 to 15 post-operative years and reported favourable outcome (defined using Crandall's criteria which is equivalent to our definition of favourable outcome of Engel classes 1 and 2) for 62% of patients. As follow-up for these patients was up to 15 years the results are similar to our findings as 66.7% of our neocorticectomy with 15 years or less follow-up had a favourable outcome.
Table 2 compares our results with similar studies.
Table 2. Comparison of % patients seizure free following resective surgery for epilepsy between different studies.
| Year after surgery | Our results (n=199) | Salanova et al. (Ref. 9) (n=145) | Rougier et al. (Ref. 18) (n=100) | Elwes et al. (Ref. 19) (n=102) | Jutila et al. (Ref. 20) (n=140) | Asztely et al. (Ref. 8) (n=65) |
|---|---|---|---|---|---|---|
| 1 | 59.8% | 66% | 66% | 56% | 57% | – |
| 2 | 56.6% | 63% | 68% | 60% | – | 54% |
| 5 | 41.4% | 60% | 51% | 67% | – | – |
| 10 | 44% | 55% | – | – | – | – |
| 15 | 25% | – | – | – | – | – |
| >15 | 31.3% | – | – | – | – | – |
There was no significant difference between our results and those of the other studies at the first and second post-operative years. Five years after surgery however, with regard to seizures, two large scale studies9, 19 had a significantly better outcome than our study (60% and 67% seizure free, respectively, compared to our 41.4%). This may have been due to inclusion in our study of the neocorticectomy patients who generally had poorer outcomes than our other patients. Our results were re-analyzed after exclusion of the neocorticectomy patients. Resultant figures are 61.4% seizure free at 2 years and 48.7% at 5 years.
Our results support the long-term efficacy of resective surgery for epilepsy, but also show (Fig. 2) that with longer follow-up duration fewer patients are classed as seizure free. Thirty-four patients (17%) experienced a late relapse of seizures which may suggest that new epileptogenic pathways may form over time. Surgery in many instances may be a mere interruption in the epileptic network of electrical spread, rather than a resection of a discreet epileptic focus, and the margins of a surgical resection may in time take over the role of the generator or peak of an epileptic network. Another possibility is that during epileptogenesis, whenever that happens, several potential sites of epileptic generation may be initiated. One focus becomes dominant, and following resection gives some years of freedom, before a secondary site then takes over.
Only three (15%) of the late relapse patients in our study with at least 1-year follow-up after relapse (3–17 years, mean 9.25) regained their seizure free status and were classed as such at last follow-up. This is in contrast to two other studies21, 22 which reported that 87% and 53%, respectively, of their late relapse patients regained their seizure free status at end of follow-up. The former study included patients who relapsed after one seizure free post-operative year, and the latter included those who relapsed after five seizure free years. Therefore, those patients who relapse later rather than earlier may not have such a good prognosis for seizure outcome. The mean duration of follow-up after relapse in the former study21 was 3.7 years and in the latter study22 information regarding duration of follow-up after relapse was not available. In our study the mean duration length of follow-up was 8.7 years. Our results are more consistent with another study23 which reported that all of their late relapse patients became refractory to medical treatment. However this study also reported that late recurrences are seen predominantly in MTS patients; in our study less than one quarter of our late relapse patients had a diagnosis of MTS.
Of those with an unfavourable end of follow-up outcome at 15 years or more, 60% had been classed as unfavourable at each stage of follow-up and 27% were late relapsers. Similarly, 71% of those with a favourable outcome at last follow-up had been classed as favourable for each stage of follow-up, 29% of which were seizure free for the entire duration of follow-up. Therefore, previous post-operative seizure history is a very good predictor of long-term outcome. Of patients who relapsed in our study, 80.7% had done so by the second year after surgery and similar findings have been demonstrated by other authors.10, 19 Patients, therefore, who have remained seizure free by the second anniversary have a very good chance of a long-term favourable outcome.
People with epilepsy have an increased mortality rate compared with the general population.24 Thirteen patients in our study died during follow-up (6.5%) and this figure is similar to results from other studies that range from 3.4% to 10.9%.5, 8, 9, 10, 18, 20, 25, 26, 27 Unfortunately cause of death could only be ascertained for 5 patients. Two patients died by suicide (1%) which is similar to another study reporting a post-operative suicide rate of 1.6%.25
Resective epilepsy surgery has been performed in Ireland since 1975. Despite the difficulties in obtaining outcome information, this series has shown that the long-term seizure outcome for Irish patients who underwent surgery for refractory epilepsy compares favourably with results from other countries. This study also allowed for the comparison of long-term outcome from an operation largely unique to this series – the neocorticectomy –with more typical methods of surgery adopted for epilepsy. Finally, this study indicates the presence of a late relapse phenomenon seen in other long-term outcome series and the poor prognosis that is associated with it.
Conflicts of interest
None.
Funding
None.
References
- . A randomized controlled trial of surgery for temporal lobe epilepsy. N Engl J Med. 2001;345:311–318
- Temporal neocorticectomy in management of intractable epilepsy: long-term outcome and predictive factors. Epilepsia. 1992;33:852–861
- . Outcome with respect to epileptic seizures. In: Engel J editors. Surgical treatment of the epilepsies. 2nd edn.. New York: Raven Press; 1993;p. 609–621
- . Protocols for surgery for epilepsy in different centres: Richmond Institute for Neurology and Neurosurgery, Dublin, Ireland. In: Luders H editors. Epilepsy Surgery. 2nd edn.. Raven Press; 1993;
- . Temporal lobectomy: long-term seizure outcome, late recurrence and risks for seizure recurrence. Brain. 2004;127(Pt 9):2018–2030
- Survival analysis of the surgical outcome of temporal lobe epilepsy due to hippocampal sclerosis. Epilepsia. 2004;45:1383–1391
- Seizure outcome after temporal lobectomy for temporal lobe epilepsy. Neurology. 2000;54(3):630–634
- . Long term follow-up of the first 70 operated adults in the Goteborg epilepsy surgery series with respect to seizures, psychosocial outcome and use of anti-epileptic drugs. J Neurol Neurosurg Psychiatry. 2007;78:605–609
- . Longitudinal follow-up of 145 patients with medically refractory temporal lobe epilepsy treated surgically between 1984 and 1995. Epilepsia. 1999;40:1417–1423
- . Temporal lobectomy for refractory epilepsy. JAMA. 1996;14(276):470–475
- Preoperative MRI predicts outcome of temporal lobectomy: an actuarial analysis. Neurology. 1995;45:1358–1363
- . Prognostic factors in the surgical treatment of temporal lobe epileptics. Neurology. 1968;18:717–731
- . Predictors of outcome of epilepsy surgery: multivariate analysis with validation. Epilepsia. 1996;37:814–821
- Temporal lobe hypometabolism on PET: predictor of seizure control after temporal lobectomy. Neurology. 1993;43:1088–1092
- Surgical outcome of patients with mesial temporal lobe epilepsy related to hippocampal sclerosis. Epilepsia. 2008;4:696–699
- . Long term seizure outcome after mesial temporal lobe epilepsy surgery: corticalamygdalohippocampectomy versus selective amygdalohippocampectomy. J Neurosurg. 2008;108(3):517–524
- Seizure and memory outcome following temporal lobe surgery: selective compared with non-selective approaches for hippocampal sclerosis. J Neurosurg. 2006;104(1):70–78
- . A longitudinal assessment of seizure outcome and overall benefit from 100 cortectomies for epilepsy. J Neurol Neurosurg Psychiatry. 1992;55:762–767
- . Outcome following resective surgery for temporal lobe epilepsy: a prospective follow-up study of 102 consecutive cases. J Neurol Neurosurg Psychiatry. 1991;54(11):949–952
- Long term outcome of temporal lobe epilepsy surgery: analyses of 140 consecutive patients. J Neurol Neurosurg Psychiatry. 2002;73:486–494
- . Late seizures in patients initially seizure free after epilepsy surgery. Epilepsia. 2006;47(3):567–573
- Prognosis after late relapse following epilepsy surgery. Epilepsy Res. 2008;78:77–81
- . Failed surgery for epilepsy—a study of persistence and recurrence of seizures following temporal resection. Brain. 2000;123:2445–2466
- . Epilepsy: frequency, causes and consequences. New York: Demos Publications; 1990;
- . Surgical treatment for partial epilepsy among Norwegian children and adolescents. Epilepsia. 1994;35(3):554–565
- . Postoperative seizure control and anti-epileptic drug use in paediatric epilepsy surgery patients: the UCLA experience, 1986–1997. Epilepsia. 1999;40(12):1740–1749
- . Temporal lobe epilepsy surgery: outcome, complications and late mortality rate in 215 patients. Epilepsia. 2002;43(2):170–174
PII: S1059-1311(10)00055-5
doi:10.1016/j.seizure.2010.03.001
© 2010 British Epilepsy Association. Published by Elsevier Inc. All rights reserved.
