Distribution of Epilepsy Syndromes in a Cohort of Children Prospectively Monitored from the Time of Their First Unprovoked Seizure

Epilepsiu, 4( ):378-383, 999 Lippincott Williams & Wilkins, Inc., Philadelphia International League Against Epilepsy Clinical Research Distribution of Epilepsy Syndromes in a Cohort of Children Prospectively Monitored from the Time of Their First Unprovoked Seizure *t Shlorno Shinnar, *$ Christine O'Dell, and "Anne T. Berg Departments of *Neurology,?Pediatrics, and $Nursing and the $Comprehensive Epilepsy Management Center, Montefiore Medical Center, The Albert Einstein College of Medicine, Bronx, New York; and "Department of Biological Sciences, Northern Illinois University, DeKalb, Illinois, U.S.A. Summary: Purpose: To assess the distribution of epilepsy syndromes and their stability in children. Methods: A cohort of 47 children with a first unprovoked seizure was prospectively recruited and followed up for a mean of 9.4 years. Etiology and epilepsy syndromes were classified by using the International League Against Epilepsy (ILAE) guidelines in the 8 children with two or more seizures. Classification was done both at time of second seizure and at last follow-up. Two-year terminal remission also was analyzed by etiology and epilepsy syndrome. Results: Etiology of epilepsy syndromes was idiopathic in 45 (5%), cryptogenic in 89 (49%), and remote symptomatic in 48 (6%). In the initial classification, 4 (63%) children had a localization-based epilepsy syndrome including idiopathic in 6, cryptogenic in 34, and symptomatic based on localization or etiology in 54. Twenty-one (%) children had a generalized epilepsy syndrome, including 9 with primary generalized epi- lepsy. Forty-seven (6%) cases were in the category of undetermined if focal or generalized. At last follow-up, there was a change in either etiology (n = 6) or the final epilepsy syndrome classification (n = 33) or both (n = 5) in 34 (9%) cases. At time of last follow-up, 44 (79%) of the children with epilepsy were in -year terminal remission, and 8 (59%) were in -year terminal remission without medications. Factors associated with a favorable prognosis included an idiopathic or cryptogenic etiology and having a localization-based idiopathic epilepsy syndrome. Conclusions: After two seizures, childhood-onset epilepsy can be classified by etiology and epilepsy syndrome. Prognosis is favorable in the majority of cases. However, the apparent syndrome may change with longer follow-up. The ability to classify these cases early in the clinical course is important if they are to be used for prognostic purposes. Key Words: Epilepsy syndrome-prognosis-epidemiology. The new International League Against Epilepsy (ILAE) classification schemes for etiology (I), seizure type (), and epilepsy syndrome (3,4) provide valuable guidelines for reporting the incidence and prognosis of childhood-onset seizures in a way that distinguishes among the different types of epilepsy with their predictably different prognoses. However, few epidemiologic studies of childhood seizures have used this classification (5-). We have followed a cohort of 47 children Accepted March 6, 999. Address correspondence and reprint requests to Dr. S. Shinnar at Epilepsy Management Center, Montefiore Medical Center, E. th Street, Bronx, NY 467, U.S.A. Presented in part at the American Epilepsy Society Meetings, San Diego, CA, December 998. Dr. Shinnar is a Martin and Emily Fisher Fellow in Neurology and Pediatrics at the Albert Einstein College of Medicine. identified at the time of their first unprovoked seizure for a mean of 9.4 years ( ). We report on the distribution of epilepsy syndromes and its relation to outcome in the 8 children who have had two or more seizures and thus meet the ILAE criteria for epilepsy (). MATERIALS AND METHODS In a prospective cohort study, we recruited and enrolled 47 children with a first unprovoked afebrile seizure seen at Montefiore Medical Center, Jacobi Medical Center, North Central Bronx Hospital, or the private practices of the Pediatric Neurology division at the Albert Einstein College of Medicine between October 983 and August 99. These children were then prospectively followed up. Eligible candidates for the study were children, age month to 9 years, with their first unpro- 378

EPILEPSY SYNDROMES IN CHILDREN 379 voked afebrile seizure. Details of the inclusion and exclusion criteria for this cohort as well of the initial evaluation have been previously reported ( ). At the time of the initial visit, informed consent was obtained from the parent and assent from the child when appropriate. Details were collected about the exact nature of the seizure including seizure characteristics, duration, number of seizures in 4 h, and any treatment given. Additional history regarding prior provoked seizures, prior neurologic insults, birth history, and family history of seizures also was coded. A physical and neurologic examination was performed on all children. An electroencephalogram was scheduled for all patients. Further laboratory examinations such as a computed tomography (CT) scan or magnetic resonance imaging (MRI) were performed when clinically indicated. After enrollment, the children were followed up by telephone interviews at least every 3 months for ascertainment of any seizure recurrence. Formal neurologic assessments were done when clinically indicated. In those children with a recurrence, records of any emergency medical care were reviewed, and the children were seen for reevaluation. A recurrence was defined as any unprovoked seizure occurring >4 h after the first seizure. The mean follow-up period after the first seizure was 9.4 years. As of September, 998, 8 (45%) children experienced a seizure recurrence and therefore meet the criteria for the diagnosis of epilepsy (). They have been followed up for a mean of 8. years after their second seizure. Of these 8 children, 8 (99%) have been followed up for years after the second seizure, and 74 (96%) have been followed up for 5 years after the second seizure. All 8 cases were reviewed by an epileptologist (S.S.), and the etiology and epilepsy syndrome classified in accordance with the ILAE guidelines (-3), based on the information available at the time of the second seizure. They were then also reclassified based on the information available at the time of last follow-up. The classification at the time of second seizure was done with blinding to subsequent outcome, including seizure type and frequency and medication response, as well as subsequent EEGs and imaging studies. The classification at final follow-up was based on all available information from the medical records, subsequent EEG and imaging studies, and follow-up telephone interviews with the subjects and their families to clarify any details that may have been unclear from the medical record. The final classification was therefore not blinded. EEGs were performed on 38 (94%) of the 47 subjects and in 78 (98%) of the 8 children with two or more seizures. EEGs were usually done after the first seizure and not repeated after the second one unless there was a clinical indication. Whereas there was no defined protocol for the EEGs, whenever possible, both awake and asleep tracings were obtained. The details on the initial EEGs in this cohort have been previously reported (). Of the 8 children with epilepsy, 7 (64%) had two or more EEGs including 6 (34%) with two EEGs, 4 (3%) with three EEGs, (6%) with four EEGs, and four (%) with five or more EEGs. Etiology was classified in accordance with the ILAE guidelines for epidemiologic studies (). Seizure type and epilepsy syndromes were classified in accordance with the ILAE criteria (-4). The only exception was that under the primary generalized epilepsies, the syndrome of random grand mal, which has been shown to be a separate syndrome from grand ma on awakening, was added to the classification as a separate entity (3). For this analysis, remission is defined as -year terminal remission with or without medications, and remission without medications is defined as -year terminal remission without medications (I, 4). For statistical analyses in this descriptive article, the following tests were used: Pearson s x test, with Yates correction when appropriate, and Fisher s Exact test (two-tail) (5). The analyses were carried out by means of BMDP (6) and SAS (7) statistical software. For all the analysis performed, a p value of <.5 was considered statistically significant. All p values are two-tailed. The study design was approved by the Institutional Review Board for the protection of Human Subjects of Montefiore Medical Center and the Albert Einstein College of Medicine. RESULTS Initial syndrome The distribution of epilepsy syndromes based on the classification at the time of the second seizure is shown in Tables and. Of the 8 children with epilepsy, etiology was idiopathic in 45 (5%), ci-yptogenic in 89 (49%), and remote symptomatic in 48 (6%). The distribution of etiologies is different from that in the entire cohort of 47 children classified based on the first seizure, in which etiology is idiopathic in 7 (7%), cryptogenic in 7 (67%), and remote symptomatic in 65 (6%) (p <.). This is due primarily to the higher recurrence risks associated with a remote symptomatic etiology and with an abnormal EEG, which is a necessary criterion for the diagnosis of an idiopathic etiology. In a few cases, additional information from the second seizure, a repeated EEG, or a new imaging study provided information after the second seizure that was not available at the time of the first seizure and altered the etiology. More than 6% of cases of epilepsy are localization related. Benign rolandic epilepsy accounts for -3% of the cases. Among the 54 localization-related cases clas- Epilepsia, Vol. 4, No,. 999

38 S. SHINNAR ET AL. TABLE. Distribution of epilepsy syndromes and outcomes in 8 children with two or more seizures Epilepsy syndrome Localization related Occipital Symptomatic based on localization or etiology Temporal lobe Frontal lobe Occipital lobe Other lobe Not localizableb Generalized epilepsies Juvenile absence Juvenile myoclonic Grand ma on awakening Random grand ma Other primary generalized and/or symptomatic syndrome if focal or generalized' Total -year remission" At onset Last follow-up All Without meds 4 (63%) 6 (4%) 4 (3%) (%) 54 (3%) 6 (3%) 9 (6%) 34 (9%) (%) 9 (%) (7%) 4 (%) (I%) 47 (6%) 8 (66%) 7 (4%) 3 (3%) 4 (%) 6 (34%) (6%) 8 (4%) 8 (5%) 33 (9%) 8 (5%) (%) (%) (%) 7 (4%) 7 (4%) 33 (8%) 8 93 (8%) 6 (%) 4 (%) ( %) 39 (7%) 9 ( %) 4 (67%) 3 (6%) 3 (6%) (69%) 8 (8%) 5 (7%) 3 (68%) 7 (58%) 4 ( %) (%) 36 (77%) 44 (79%) 7 (63%) 5 (96%) 3 (96%) (%) 4 (44%) 7 (78%) (33%) 3 (6%) (4%) (34%) 3 (68%) 6 (9%) 5 (6%] I (8%) 3 (75%) I(5%) 3 (64%) 8 (59%) Refers to -year terminal remission based on initial syndrome classification. All refers to remission with or without medications, and without meds refers to remission without medications. 'I Symptomatic localization related by virtue of etiology only. Children with generalized tonic-clonic seizures and a normal EEG were classified in this category. sified as symptomatic by virtue etiology or localization, 5 could be localized, with the most common localization being temporal lobe (Table ). Approximately one fourth of the cases are classified as undetermined, whether focal or generalized. In this category belong the children with a tonic-clonic seizure and a normal EEG. This is a much smaller proportion than the 69 (4%) of the original cohort of 47 children with a first seizure who were in this category, but is still a substantial proportion of cases. Of (%) children with a generalized epilepsy, 9 had a primary generalized epilepsy, with the most common being random grand ma (n = ): Note that in principle, syndromes such as juvenile myoclonic epilepsy, childhood absence, and juvenile absence will not enter this study. However, in a few cases, either the presentation was atypical or the history of myoclonic jerks or staring was not initially obtained despite our best efforts, and these cases remained in the cohort. Final syndrome The distribution of epilepsy syndromes based on the classification at the time of last follow-up is shown in Tables and. There was a change in either etiology, final epilepsy syndrome classification, or both in 34 TABLE. Distribution of epilepsy syndromes and outcomes in 8 children with two or more seizures by etiology -year remission" Epilepsy syndrome At onset Last follow-up All Without meds Primary generalized Generalized if focal or generalizedh Remote symptomatic Generalized if focal or generalizedb Total 45 (5%) 6 (4%) 9 (%) 89 (49%) 5 (7%) 38 (%) 48 (6%) 38 (%) I8 48 (6%) 7 (5%) (%) 8 (45%) 5 (9%) 9 (6%) 53 (9%) 4 (3%) 7 (4%) 4 (%) 8 39 (87%) 6 ( %) 3 (68%) 7 (79%) 4 (8%) 9 (76%) 35 (73%) 7 (7%) 7 (77%) 44 (79%) 3 (67%) 5 (96%) 5 (6%) 6 (69%) 33 (66%) 7 (7%) 7 (35%) 4 (37%) 3 (33%) 8 (59%) a Refers to -year terminal remission based on initial syndrome classification. All refers to remission with or without medications, and without meds refers to remission without medications. Children with generalized tonic-clonic seizures and a normal EEG were classified in this category. Epilepsia, Val. 4, No., 999

EPILEPSY SYNDROMES IN CHILDREN 38 (9%) of cases (Table 3). Of the 8 cases, the final etiology was different from that initially assigned in 6 (9%) cases. Three cases initially were classified as idiopathic and then reclassified as cryptogenic. Twelve children were initially classified as cryptogenic and subsequently classified as either idiopathic (n = 6) or remote symptomatic (n = 6) based on subsequent EEGs, imaging data, or clinical information. The final syndrome was different from the initial syndrome assignment in 33 (8%) cases. This included cases whose epilepsy was initially classified as undetermined, whether focal or generalized, and subsequently reclassified as partial (n = ) or generalized (n = ) syndromes, based on additional information obtained during the follow-up. Four cases initially classified as benign rolandic epilepsy were subsequently reclassified as benign occipital (n = ), symptomatic localizationbased epilepsy of cryptogenic etiology (n = l), and juvenile absence (n = ). The last case initially was seen at age 4 years with seizures and an EEG characteristic of benign rolandic epilepsy. She then went into remission. In adolescence, she developed absence seizures, and her EEG then showed 3-Hz spike and wave consistent with her new clinical syndrome of juvenile absence epilepsy. Six cases initially diagnosed as cryptogenic partial epilepsy were then reclassified as having a localizationrelated epilepsy, symptomatic by virtue of localization, on the basis of subsequent EEGs, although the etiology remained cryptogenic. In addition, six children, five of whom had a remote symptomatic etiology, initially diagnosed as a partial epilepsy or as unclear, whether focal or generalized, were later reclassified as having a sec- ondarily generalized epilepsy syndrome (Lennox- Gastaut) based on the clinical evolution of the seizure disorder and a subsequent EEG demonstrating generalized slow spike and wave. Cases with many seizures were more likely to be reclassified. Eighteen (53%) of 34 children whose etiology or syndromes were reclassified had seizures compared with 9 (%) of 48 children in whom no change in classification occurred (p <,). Remission As of last follow-up on September, 998, 44 (8%) of the 8 patients on whom years of follow-up after the second seizure was available were in -year terminal remission, and 8 (6%) were in -year terminal remission without medications. Of the 74 cases for whom 5 years of follow-up are available, 8 were in 5-year terminal remission, and 83 (48%) in 5-year terminal remission without medications. The distribution of outcomes by etiology and epilepsy syndromes is shown in Tables I and. Two-year terminal remission was a function of both etiology and the specific epilepsy syndrome (Tables and ). Remission rates were similar for all etiologies. However, remission without medications was more likely in children with idiopathic (67%) or cryptogenic (69%) etiology than in those with remote symptomatic etiology (35%) (p <.). Among children with idiopathic seizures, those with localization-related epilepsies had a higher rate of remission (p =.3) and of remission without medications (p <.) than children with primary generalized idiopathic epilepsy syndromes (Tables and ). All children initially diagnosed Initial classification TABLE 3. Change in syndrome or etiology from initial presentation Final classification No. Etiology Syndrome Etiology Syndrome Comment I 6 3 3 Random grand ma Random grand ma Temporal lobe Generalized Occipital Juvenile absence, remote symptomatic; undetermined, undetermined if focal or generalized Juvenile myoclonic Temporal lobe epilepsy Other primary generalized Other primary generalized Specific localization related Symptomatic generalized True change with new seizure type after remission Change in seizure type Seizure type changed Clinical picture changed or abnormal subsequent imaging Clinical picture changed Change in seizure type Evolution of seizure types Epilepsiu, Vol. 4, No., 999

38 S. SHINNAR ET AL. with an idiopathic localization-related epilepsy had a favorable outcome (Tables and ), including the four cases for whom the final syndrome classification was different from the initial one. Once idiopathic syndromes are excluded, there were no significant differences in outcomes, which were generally favorable, between the different localization-related syndromes. DISCUSSION This is a study of the long-term outcomes of a cohort of 47 children followed up from the time of their first unprovoked seizure, of whom 8 have met the criteria for epilepsy to date. The strengths of this study are the long-term follow-up with a high retention rate and the detailed information available, permitting accurate classification of seizure type, epilepsy syndrome, and etiology, as well as the seizure outcomes. The analysis of recurrence risk in this cohort ( ) and the initial EEG () and imaging (8) data have been previously reported. Few population- or community-based series have had sufficient data to classify the patients by epilepsy syndromes (5-,9). Three recent studies of newly diagnosed childhood epilepsy from Connecticut (5), the Netherlands (6), and Finland (9,lO) classified the epilepsy syndromes in accordance with the ILAE guidelines (3,4). The distribution of epilepsy syndromes in this cohort is necessarily different from that reported in those studies, because children with childhood absence and various myoclonic epilepsies will be excluded from this study as they invariably have had more than one event by the time of initial presentation to medical attention. Similarly the distribution of partial epilepsies will be skewed away from temporal lobe epilepsy, which tends receive medical attention after several events have already taken place. This accounts for the relatively high proportiog of cases in this cohort with benign rolandic epilepsy and with random grand mal, both of which tend to be seen after a single convulsive episode. The overall prognosis of this cohort is favorable. At last follow-up, three fourths of those with epilepsy were in -year terminal remission, and almost half were in 5-year terminal remission. As the other 5 children, who have only had one seizure, are by definition in terminal remission at last follow-up, the numbers for the entire cohort of 47 are even more favorable. Not surprisingly, etiology and the specific epilepsy syndrome are associated with a differential prognosis, which is consistent with other published studies (4-,4,9-6). Most of this cohort has had a relatively benign course, with only 47 (6%) children having had seizures to date. Even the remote symptomatic cases in this cohort have had, with a few notable exceptions, a relatively benign course. There is increasing evidence that long- term prognosis is more a function of the underlying syndrome than of the number of seizures or of the antiepileptic drug (AED) therapy used (5-8). The favorable prognosis of this cohort of children is attributable to the distribution of epilepsy syndromes that are seen with a first unprovoked seizure. Of interest and concern is the number of cases for which the etiology or the specific epilepsy syndrome changed with time. Etiologic classification will change based on a syndrome change or based on additional information from subsequent EEGs or imaging data. The fact that with time there were fewer cases with cryptogenic etiology or with a relatively nonspecific syndrome classifications, such as unclear whether focal or generalized or cryptogenic partial seizures, is not surprising, as those with ongoing seizures will tend to have multiple EEG studies and are more likely to have good imaging studies, which will further elucidate the underlying syndrome (9-3). However, a surprising number of cases met the criteria for a specific epilepsy syndrome such as benign rolandic epilepsy or a specific primary generalized syndrome and then changed with time. Reassuringly, in some cases, the change was relatively minor, such as from benign rolandic to benign occipital, and all cases initially classified as having an idiopathic localization-related epilepsy have had a benign clinical course. The evolution of some cases with onset as focal seizures into secondarily generalized syndromes is also well established (3), although it is rather surprising to see it in this population. Two recent abstracts also suggested that the initial classification of epilepsy syndromes may change after a few years of follow-up in % (33) and 5% (34) of the subjects. These data suggest that syndrome classification at the very onset of the epilepsy may not be as robust and reliable as when the classification is performed after the syndrome has either evolved or remained stable for some time. More data on this issue from prospective studies that have followed up and classified patients from onset are needed, as syndromes are widely used in predicting long-term outcome (3,4), and this presumes that they will remain stable. CONCLUSIONS After two seizures, cases of childhood-onset epilepsy can be classified according to the ILAE guidelines for classification of etiology, seizure types, and epilepsy syndromes. However, the classification may change with longer follow-up as the epilepsy syndrome evolves. The ability to classify these cases is important for a better understanding of the long-term outcomes of these children, as the prognosis of the various syndromes is quite different. More data from prospective studies are needed on the stability of the ILAE syndrome classification over time. Epilepsia, Vnl. 4, NO. lo, 999

EPILEPSY SYNDROMES IN CHILDREN 383 Acknowledgment: We thank Ms. Marta Alemany for helping maintain this cohort in follow-up, Myint Maw, M.D., M.P.H., for his help with statistical analysis, and Avraham Shinnar for data entry and computer programming. We are indebted to all the house staff and attending physicians who allowed us to recruit their patients into the study. We also acknowledge the cooperation of the New York City Health and Hospitals Corporation and Jacobi Hospital and North Central Bronx Hospital. This study was supported in part by grant R NS6.5 (S. Shinnar) from the National Institute of Neurological Disorders and Stroke, NIH, Bethesda, MD. REFERENCES. Commission on Epidemiology and Prognosis of the ILAE. Guidelines for epidemiological studies on epilepsy. Epilepsia 993;34: 59-6.. Commission on Classification and Terminology of the ILAE. Proposal for revised clinical and electroencephalographic classificas. Epilepsia 98;:489-5. cation and Terminology of the ILAE. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 989;3:389-99. 4. Roger J, Bureau M, Dravet Ch, Dreifuss FE, Perret A, Wolf P, eds. Epileptic syndromes in infancy, childhood and adolescence. nd ed. London: John Lihbey, 99: 48. 5. Berg AT, Shinnar S, Levy SR, Testa FM. Newly diagnosed epilepsy in children: presentation at diagnosis. Epilepsia 999;4: 445-5. 6. Callenbach PMC, Geerts AT, Arts WFM, et al. Familial occurrence of epilepsy in children with newly diagnosed multiple seizures: Dutch study of epilepsy in childhood. Epilepsia 998;39: 33-6. 7. Loiseau J, Loiseau P, Guyot M, Duche B, Dartigues JF, Aublet B. Survey of seizure disorders in the French southwest: I. incidence of epileptic syndromes. Epilepsiu 99;3 :39-6. 8. Manford M, Hart YM, Sander J, Shorvon SD. The National General Practice Study of Epilepsy: the syndromic classification of the International League Against Epilepsy applied to epilepsy in a general population. Arch Neurol 99;49:8-8. 9. Sillanpaa M, Jalava M, Kaleva, Shinnar S. Long-term prognosis of seizures with onset in childhood. NEngl JMed 998;338:75-.. Sillanpaa M, Jalava M, Shinnar S. Epilepsy syndromes in population based cohort of 45 patients with childhood-onset seizures in Finland. Pediatr Neurol (in press).. Shinnar S, Berg AT, Moshe SL, et al. The risk of seizure recurrence following a first unprovoked afebrile seizure in childhood: an extended follow-up. Pediatrics 996;98: 6-5.. Shinnar S, Kang H, Berg AT, Goldensohn ES, Hauser WA, Moshe SL. EEG abnormalities in children with a first unprovoked seizure. Epilepsia 994;35:47-6. 3. Greenberg DA, Durner M, Resor S, Rosenbaum D, Shinnar S. The genetics of idiopathic generalized epilepsies of adolescent onset: differences between juvenile myoclonic epilepsy and epilepsy with random grand ma and with awakening grand mal. Neurology 995;45:94-6. 4. Annegers JF, Hauser WA, Elveback LR. Remission of seizures and relapse in patients with epilepsy. Epilepsia 979;:79-37. IS. Kelsey JL, Thompson WD, Evans AS. Methods in observational epidemiology. New York: Oxford University Press, 986. 6. Dixon W. BMDP statistical software. Berkeley: University of California Press, 99. 7. SAS Institute Inc. SAS user s guide statistics. 985 ed. Cary, NC: SAS Institute, Inc., 985. 8. O Dell C, Shinnar S, Mitnick R, Berg AT, Moshe SL. Neuroimaging abnormalities in children with a first afehrile seizure. Epilepsia 997;38(suppl 8): 84. 9. Berg AT, Hauser WA, Shinnar S. The prognosis of childhoodonset epilepsy. In: Shinnar S, Amir N, Branski D, eds. Childhood seizures. Basel: S. Karger, 995:93-9.. Sofijanov NG. Clinical evolution and prognosis of childhood epilepsies. Epilepsia 98;3:6-9.. Brorson LO, Wranne L. Long-term prognosis in childhood epilepsy: survival and seizure prognosis. Epilepsia 987;8:34-3.. Oka E, Yamatogi Y, Ohtsuka, Ohtahara S. Clinical course and prognosis of childhood epilepsy. Acta Paediatr Jpn 989;3 :59-66. 3. Berg AT, Levy SR, Novotny EJ, Shinnar S. Predictors of intractable epilepsy in childhood: a case control study. Epilepsia 996; 37:4-3. 4. Hauser WA, Hesdorffer DC. Epilepsy: frequency, causes and consequences. New York: Demos, 99. 5. Sander JWAS. Some aspects of prognosis in the epilepsies: a review. Epilepsia 993;34: 7-6. 6. Cockerell OC, Johnson AL, Sander JWAS, Hart YM, Shorvon SD. Remission of epilepsy: results from the National General Practice Study of Epilepsy. Lancet 995;346: 44. 7. Shinnar S, Berg AT. Does antiepileptic drug therapy prevent the development of chronic epilepsy. Epilepsia 996;37:7-8. 8. Berg AT, Shinnar S. Do seizures beget seizures? An assessment of the clinical evidence in humans. J Clin Neurophysiol 997;4:-. 9. Bronen RA. Epilepsy: the role of MR imaging. AJR Am J Roentgenol 99;59:65-74. 3. Yatishita A, Arai N, Machara T, Shimizu H, Tokumara AM, Oda M. Focal cortical dysplasia: appearance on MR images. Radiology 997;3:553-9. 3 I. Gunay M. Neuronal migration disorder presenting with mild clinical symptoms. Pediatr Neurol 996;4:534. 3. Ohtsuka Y, Ohmori I, Oka E. Long-term follow-up of childhood epilepsy associated with tuberous sclerosis. Epilepsia 998;39: 58-63. 33. Arts WF, Geerts A, Brouwer, Peters A, Stroink H, van Donselaar C. Classification schemes in childhood epilepsy: reliability and causes of discrepancy. Epilepsia 997;38(suppl 3):. 34. Loiseau J, Picot MC, Jallon P, Dartigues JF, Loiseau P. Classification and incidence of epileptic syndromes in a prospective study: reliability and causes of change. Epilepsia 998;39(suppl 6): 8. Epilepsia, Vol, 4, No., 999