Is benign rolandic epilepsy genetically determined?

Benign rolandic epilepsy (BRE) is considered to be a genetically determined idiopathic partial epilepsy. We studied twins with BRE and compared the concordance with a twin sample of idiopathic generalized epilepsy (IGE). All eight BRE pairs (six monozygous [MZ], two dizygous [DZ]) were discordant. MZ pairwise concordance was 0 (95% confidence interval [CI], 0-0.4) for BRE compared with 0.7 (95% CI, 0.5-0.9) for 26 IGE MZ pairs. Our data suggest that conventional genetic influences in BRE are considerably less than for IGE, and other mechanisms need to be explored.     

Epilepsies in twins: Genetics of the major epilepsy syndromes

We studied twins to examine the genetics of epilepsy syndromes. We ascertained 358 twin pairs in whom one or both reported seizures. After evaluation, 253 of 358 (71%) had seizure disorders and 105 pairs were false positives. Among the monozygous (MZ) pairs, more were concordant for seizures (48 of 108; casewise concordance = 0.62 ± 0.05) than among the dizygous (DZ) pairs (14 of 145; casewise concordance = 0.18 ± 0.04). In 94% of concordant MZ pairs, and 71% of concordant DZ pairs, both twins had the same major epilepsy syndrome. When analyzed according to major epilepsy syndrome, the casewise concordances for generalized (MZ = 0.82; DZ = 0.26), both idiopathic (MZ = 0.76; DZ = 0.33) and symptomatic (MZ = 0.83; DZ = 0), were greater than those for partial epilepsies (MZ = 0.36; DZ = 0.05), with intermediate values seen for febrile seizures (MZ = 0.58; DZ = 0.14) an dunclassified epilepsies (MZ = 0.53; DZ = 0.18). We conclude that genetic factors are particularly important in the generalized epilepsies but also play a role in the partial epilepsies. The high frequency of concordant MZ pairs with the same major syndrome strongly suggets there are syndrome-specific genetic determinants rather than a broad genetic predisposition to seizures.     

Familial temporal lobe epilepsy: A common disorder identified in twins

We describe a new syndrome of familial temporal lobe epilepsy in 38 individuals from 13 unrelated white families. The disorder was first identified in 5 concordant monozygotic twin pairs as part of a large-scale twin study of epilepsy. When idiopathic partial epilepsy syndromes were excluded, the 5 pairs accounted for 23% of monozygotic pairs with partial epilepsies, and 38% of monozygotic pairs with partial epilepsy and no known etiology. Seizure onset for twin and nontwin subjects usually occurred during adolescence or early adult life. Seizure types were simple partial seizures with psychic or autonomic symptoms, infrequent complex partial seizures, and rare secondarily generalized seizures. Electroencephalograms revealed sparse focal temporal interictal epileptiform discharges in 22% of subjects. Magnetic resonance images appeared normal. Nine affected family members (24%) had not been diagnosed prior to the study. Pedigree analysis suggested autosomal dominant inheritance with age-dependent penetrance. The estimated segregation ratio was 0.3, indicating an overall penetrance of 60% assuming autosomal dominant inheritance. The mild and often subtle nature of the symptoms in some family members may account for lack of prior recognition of this common familial partial epilepsy. This disorder has similarities to the El mouse, a genetic model of temporal lobe epilepsy with a major gene on mouse chromosome 9, which is homologous with a region on human chromosome 3.     

The occurrence of epilepsy and febrile seizures in Virginian and Norwegian twins

Twin studies provide an efficient method for examining the importance of genetic and environmental factors in the etiology of disorders such as epilepsy. Population-based twin registries are especially valuable for studies of this type since effects of reporting and self-selection biases on the resulting data are minimized. Among 14,352 twin pairs contained in the Virginia and Norwegian twin panels for whom questionnaire information was available, there was a history of epilepsy in one or both members of 286 pairs; febrile seizures were reported in 257 pairs. Analyses of questionnaire data revealed no significant differences in concordance rates between Virginian and Norwegian twins for either epilepsy or febrile seizures. Probandwise concordance rates for epilepsy were 0.19 in monozygotic twins and 0.07 in dizygotic twins. Analogous rates for febrile seizures were 0.33 (monozygotic) and 0.11 (dizygotic). These results provide further evidence that genetic factors do have a role in the expression of epilepsy and febrile seizures.     

Benign Focal Epilepsies of Childhood

Until relatively recently, genetic influences in partial seizures were thought to be of minimal importance. However, with further identification of childhood benign partial seizures it is becoming clear that inheritance plays a major role in the pathogenesis of these seizures. Diagnostic criteria proposed for benign partial seizures include absence of neurologic or intellectual deficits, family history of epilepsy, onset of seizures after age 2 years, stereotyped brief seizures, frequent nocturnal occurrence, spontaneous remission in adolescence, and electroencephalograms (EEGs) demonstrating spikes with a distinctive morphology and localization superimposed on normal background activity. The two most commonly described benign partial epilepsies of childhood are benign Rolandic epilepsy (BRE) and benign occipital epilepsy (BOE). Both disorders begin in childhood, are associated with characteristic EEG patterns, have seizures that are easily controlled with medication, often are familial, and have an excellent prognosis. The other benign partial seizure disorders in children that have been described are not as well studied as BOE and BRE, and the role of inheritance pattern, if any, is less clear.     

Epileptic seizures and syndromes in twins: the importance of genetic factors

The role of genetic factors in the occurrence of epilepsy syndromes was studied in twins recruited from the population-based Danish Twin Registry. A total of 34,076 twins were screened for epilepsy. Cases were confirmed and classified by two neurologists according to the classification systems of the International League Against Epilepsy (ILAE). A total of 214 twin pairs with epileptic seizures and 190 pairs with epilepsy were ascertained. Significantly higher concordance rates were found for monozygotic (MZ) compared to dizygotic (DZ) twins for both epileptic seizures (0.56 for MZ and 0.21 for DZ pairs, P<0.001) and for epilepsy (0.49 for MZ and 0.16 for DZ pairs, P<0.001). Concordance rates were also higher for MZ twins compared to DZ twins for both generalized epilepsy (0.65 for MZ and 0.12 for DZ) and for localization-related epilepsy (0.30 for MZ and 0.10 for DZ). In twin pairs where both members had seizures, 83% of MZ and 65% of DZ pairs had the same major epilepsy syndrome. Genetic factors were found to account for 80% of the liability to both epileptic seizures and epilepsy. In conclusion, analysis of this neurologist-verified epilepsy twin data set has confirmed that genetic factors have a substantial impact on the etiology of epileptic seizures as well as on the occurrence of both generalized and partial epilepsies.     

Periventricular Nodular Heterotopia: Epileptogenic Findings

Summary: Purpose : We studied 17 patients with periventricular nodular heterotopia (PNH) to further investigate the electroclinical pictures and semiology of the associated seizures.
Methods : PNH was diagnosed by means of magnetic resonance imaging (MRI). The patients' clinical and familial histories were carefully analyzed, and their electroclinical features and course of epilepsy followed for periods ranging from 10 months to 22 years. The electroclinical data were compared with those of previously reported PNH cases.
Results : The patients were subdivided into those with bilateral (7) and unilateral (10) PNH. The former were mainly characterized by structural abnormalities in the posterior cerebral fossa and multiple seizure types; the latter were characterized by the paratrigonal location of the malformation and, frequently, by elementary seizures with a visual or auditory onset. Focal seizures were drug resistant in most cases. The interictal EEG abnormalities were always focal and consistent with the location of the PNH. A previously unreported photic driving of posterior background activity was observed in all patients and was always consistent with the PNH location.
Conclusions : Conclusions: Our present findings and previously reported data show that bilateral and unilateral PNH cases are different in their morphological and electroclinical features and may be determined by different etiologies. The female predominance, frequent familial occurrence, and positive family history for epilepsy suggest that genetic factors may be involved in the genesis of bilateral and symmetrical PNH, whereas the presence of prenatal risk factors and its location in the watershed paratrigonal area suggest that vascular mechanisms may determine unilateral PNH.     

Pattern-sensitive epilepsy: electroclinical characteristics, natural history, and delineation of the epileptic syndrome

PURPOSE: To elucidate the electroclinical features and long-term outcome of patients with pattern-sensitive epilepsy. METHODS: We reviewed the clinical and electroencephalographic (EEG) findings of 73 (43 female and 30 male) patients in whom pattern-sensitive epilepsy was diagnosed at Mayo Clinic (Rochester, Minnesota, U.S.A.) from 1950 through 1999. We contacted patients and their relatives by letter or telephone to obtain the latest seizure and quality-of-life outcomes. RESULTS: The median age at onset of seizures was 12.8 years (range, 0.6-32.9 years). Most patients had absence, myoclonic, or generalized tonic-clonic seizures. Interictal epileptiform discharges in the EEG were detected in 61 (83.6%) patients and were generalized in 54 (74%). Paroxysmal epileptiform discharges in the EEG elicited with standard patterns were all generalized in two thirds of patients but were restricted to the posterior head region in one-third. Eight (11%) patients did not exhibit photosensitivity. Television was the most common precipitant [30 patients (41%)]. Twenty-nine patients gave a clear history of one or more seizures precipitated while viewing environmental patterns such as window screens, garments, tablecloths, and ceiling tiles; the rest of the patients admitted that they preferred to avoid looking at patterned objects because these objects made them uncomfortable. The electroclinical features suggested juvenile myoclonic epilepsy in 14 patients, progressive myoclonus epilepsy in three, progressive familial cerebellar ataxia with myoclonus in two, and severe myoclonic epilepsy of infancy in one. During a median follow-up period of 15.7 years, 25 (45.5%) of 55 patients who were followed up for > or =5 years achieved complete seizure remission. The median age at remission was 24.4 years. The absence of progressive neurologic disease was correlated significantly with remission; a family history of seizures showed a trend in favor of remission. More than two thirds of the patients did not consider the seizures an impediment to their family life or to educational and occupational achievements. CONCLUSIONS: Although pattern sensitivity as a trait occurs in various epileptic syndromes, pattern-sensitive epilepsy is a readily distinguishable subtype of the visually provoked reflex epilepsies. In our opinion, the location and extent of the excitable region or regions within the visual cortex concerned with different attributes of visual function dictate susceptibility to a specific trigger (intermittent light, pattern, or color) or closely related multiple triggers and the resultant electroclinical phenomenon.     

Genetic and Environmental Contributions to Perceived Intensity and Pleasantness of Androstenone Odor: An International Twin Study

We estimated the genetic and environmental components of variation in perceived intensity and pleasantness of androstenone, an odorous compound showing specific anosmia, by modeling twin data from Finland, Denmark, the UK, and Australia. The pooled data comprised 917 twin individuals (338 are male and 579 are female; aged from 10 to 83years) including 126 complete monozygous and 264 dizygous twin pairs as well as 137 twin individuals without their co-twin. They rated intensity and pleasantness of androstenone and citronellal (control) odors using nine categories. Additive genetic effects (heritability) contributed 28 and 21% to the variation in the perceived intensity and pleasantness of androstenone, respectively, but negligibly to variations in citronellal perception. A strong genetic correlation existed between the intensity and pleasantness of androstenone, whereas the environmental correlation was negligible. These results suggest that both intensity and pleasantness of androstenone are moderately influenced by genetic factors and that the traits are modified by an overlapping set of genes.     

Severe epilepsy with multiple independent spike foci.

SUMMARY: Severe epilepsy with multiple independent spike foci is an electroclinical entity with the following characteristics: (1) EEG showing multiple independent spike foci (three or more foci in both hemisphere, i.e., at least one in each hemisphere) and diffuse slowing of the background activity, (2) very frequent multiple types of seizures but mainly generalized minor seizures, (3) frequent association with mental retardation and neurologic abnormalities, (4) underlying causes of various nonspecific prenatal, perinatal, and postnatal cerebral conditions, and (5) poor prognoses for seizures and psychomotor development. It represents a diffuse encephalopathy with mutual transition between other age-dependent epileptic encephalopathies. This electroclinical condition may be appropriately named Markand-Blume-Ohtahara syndrome, after Markand and Blume, who initially described in detail the EEG pattern of multifocal spike discharges with its major clinical correlates, and Ohtahara, who extensively studied the group of patients with both clinical and EEG characteristics and proposed the concept of severe epilepsy with multiple independent spike foci, classifying it as one type of generalized epilepsy.     

Atypical rolandic epilepsy

Typical benign rolandic epilepsy (BRE) is a frequent and well-delineated epileptic syndrome in childhood. Mild cognitive and behavioral difficulties are increasingly recognized in the course of BRE and should not be considered as atypical features. Atypical features are recognized on electroclinical grounds. These features, particularly early age at onset and frequent spikes or spike-wave discharges, seem to be risk factors for neuropsychological deficits but also for an atypical evolution of BRE. Atypical evolutions of BRE are defined by the appearance of severe neuropsychological impairments and continuous spike-and-waves during slow sleep (CSWSS). The clinical expressions of these situations correspond to the syndromes known as atypical benign focal epilepsy of childhood (ABFEC), status of BRE, Landau-Kleffner syndrome (LKS), and CSWSS syndrome, which may be part of a continuum related to BRE.     

Symptomatic occipital lobe epilepsy following neonatal hypoglycemia

This study reports on the clinical, electrophysiologic, and neuroradiologic aspects of patients with epilepsy secondary to neonatal hypoglycemia. Fifteen patients with epilepsy and/or posterior cerebral lesions, and neonatal hypoglycemia were studied in the epilepsy clinic between February 1990 and March 2003. The mean age was 12 years. The different types of neonatal hypoglycemia were as follows: four patients had transitional-adaptive, seven classic transient, two secondary-associated, and two severe recurrent hypoglycemia. As to epilepsy, we recognized a larger group of 12 patients characterized by focal seizures and posterior abnormalities on the electroencephalogram, the majority of whom had a good outcome, and a second group of two patients presenting electroclinical features of encephalopathy with refractory seizures. All patients except two manifested parieto-occipital lesions on neuroradiologic images. Neurologic examination was normal in one patient. Six patients had microcephaly; eight manifested visual disturbances. Fourteen patients were mentally retarded. One had a pervasive developmental disorder. This study indicates neonatal hypoglycemia may cause posterior cerebral lesions, abnormal findings at neurologic examination, and symptomatic epilepsy, most frequently occipital lobe epilepsy, usually with a good prognosis, and occasionally epileptic encephalopathy with refractory seizures. MRI studies are essential to define the characteristics of cerebral lesions after neonatal hypoglycemia.     

Benign Rolandic Epilepsy: High Central and Low Central Subgroups

Summary: The spikes in benign rolandic epilepsy (BRE) typically involve both the central and midtemporal regions as recorded on standard EEG montages, but the seizures are characterized by sensorimotor manifestations that are rarely referable to the temporal lobe. To study this apparent disparity, we evaluated the field distribution of interictal spikes in 33 BRE patients using closely spaced electrodes (AEEGS 1990) arranged over perisylvian cortex. None of the 33 patients showed maximum negativity in the midtemporal regions (T3/T4). Instead, maximum negativity was evident in the high central region (C3/C4) in 10 children (30.3%) and in the low central region (C5/C6) in 23 (69.7%). Hand involvement was significantly frequent (50%) in the high central group, and drooling with oromotor involvement was a distinctive symptom (65.2%) in the low central group. Our findings indicate that the spikes in patients with BRE are exclusively suprasylvian in origin and correlate with two electroclinical subgroups.     

Benign partial epilepsies of adolescence: a report of 37 new cases

PURPOSE: To delineate the electroclinical features of patients with partial seizures in adolescence with a benign outcome. METHODS: Patients were recruited in five different Italian epilepsy centers. Patients were selected among those with partial seizures between ages 11 and 17 years. We excluded benign childhood epilepsies, those with neurologic or mental deficits, and those with neuroradiologically documented lesions. We also excluded patients with less than 3 years' follow-up or who were still receiving antiepileptic therapy. RESULTS: There were 37 (22 male, 15 female) patients. Seizures started at the mean age of 14.5 years (range, 11-16.11). Two main electroclinical patterns emerged: 16 of 37 patients had somatomotor seizures frequently associated with focal theta discharges involving the centroparietal regions. Ten of 37 patients showed versive seizures and interictal spiking involving the posterior regions. A third group had clinical characteristics resembling the cases described by Loiseau. All had a favorable outcome. CONCLUSIONS: This relevant multicenter study further confirms the existence of benign partial epilepsies with onset during adolescence.     

Absence seizures in the first 3 years of life: an electroclinical study of 46 cases

Purpose: Early onset absence seizures have been considered a rare heterogeneous group with a poor prognosis. Only few patients may be categorized into well‐known syndromes. We have evaluated electroclinical features, evolution, and the nosologic boundaries of early onset absence seizures. Methods: Forty‐six neurologically normal patients with absence seizures associated with bilateral, synchronic, or asynchronic, and symmetric or asymmetric spike‐and‐wave paroxysms with onset in the first 3 years of life were included. Patients with abnormal neurologic examination and brain imaging were excluded from the study. Key Findings: In our study, 39 patients met the clinical and electroencephalography (EEG) criteria of well‐defined epileptic syndromes. Childhood absence epilepsy was found in 11 patients, benign myoclonic epilepsy in infancy in 18 patients, eyelid myoclonic epilepsy in 4, and epilepsy with myoclonic absences in 6. We did not find clinical and EEG criteria of well‐recognized epileptic syndromes in seven children. Nine of 11 patients with simple absence seizures became seizure free. All these patients had normal neurologic and neuropsychological evaluations. Of the 35 patients who had absence seizures associated with myoclonic seizures, 20 became seizure free. Fifteen of 35 children continue having seizures. At the last visit, 20 of these 35 patients had normal neurologic and neuropsychological evaluations, 11 presented with mild mental retardation, and 4 with severe mental retardation. Significance: Epilepsies with absence seizures of early onset are relatively uncommon. Most of the patients had well‐defined epileptic syndromes with a variable evolution. The evolution depended on the epileptic syndromes.     

Familial risks for epilepsy among siblings based on hospitalizations in Sweden

PURPOSE: Epilepsy is a common disabling condition, with high heritability according to twin studies. Characterization of familial risks for common subtypes of epilepsy will advance the search for the heritable causes of these conditions and their underlying mechanisms. We aim at defining familial risks for siblings to be hospitalized because of epilepsy. METHODS: A nationwide ad hoc epilepsy database was constructed by linking the Multigeneration Register on 0- to 69-year-old siblings to the Hospital Discharge Register for data on epilepsies covering the years 1987-2001. Standardized risk ratios (SIRs) were calculated for affected sibling pairs by comparing them to those whose siblings had no epilepsy. RESULTS: Among a total of 26,799 hospitalized cases, 598 affected siblings were identified with a familial SIR of 2.35; the SIR was highest at ages 0-4 years (6.82). Infantile spasms showed the highest risk for any subtype (10.45), when a co-sibling was diagnosed with any epilepsy. When both siblings were diagnosed with a concordant (same) subtype of epilepsy, the SIRs were high, i.e. 8.43 for generalized idiopathic epilepsy, 2.56 for partial epilepsy, 24.72 for status epilepticus and 24.39 for other epilepsies. Generalized idiopathic epilepsy was also associated with grand mal (4.06) and other epilepsies (7.61). The numbers of cases were small but concordant diagnoses always showing higher SIRs compared with discordant diagnoses. CONCLUSIONS: Within the limits of the present sample size, our results suggest high familial aggregation for certain subtypes of epilepsy for which distinct genetic mechanisms may underlie.     

Contribution of magnetic resonance imaging in 100 cases of refractory partial epilepsy with normal CT scans

One hundred epileptic patients were included in this study according to the following criteria: intractable partial epilepsy, normal CT scan and focal EEG abnormalities. Eighty-nine patients were suffering from complex partial seizures of temporal or frontal origin, 55 and 34 cases respectively. Eleven patients presented with only simple partial seizures. MRI was abnormal in 31 patients. The abnormalities were: focal T2 increased signal intensity (13 cases) most often temporal (10 cases), cryptic arteriovenous malformation (4 cases), focal T1 and T2 signal abnormality (4 cases), focal atrophy (2 cases) and multiple abnormal T2 signals scattered in the white matter (8 cases). The site of MRI abnormalities was consistent with electroclinical data in 22 patients, of whom 20 had a temporal lobe epilepsy. Thus MRI proved to be more often abnormal in temporal than in frontal lobe epilepsy (36 p. 100 and 5.9 p. 100 respectively) when the CT scan is normal. However MRI data, particularly focal T2 hypersignals should be confronted to electroclinical and metabolic findings whenever functional surgery is considered.     

Benign familial infantile seizures: further delineation of the syndrome

Benign familial infantile seizures are an autosomal dominant epilepsy disorder that is characterized by convulsions, with onset at age 3 to 12 months and a favorable outcome. Benign familial infantile seizures have been linked to chromosome 19q whereas infantile convulsions and choreoathetosis syndrome, in which benign familial infantile seizure is associated with paroxysmal choreoathetosis, has been linked to chromosome 16p 12-q12. Many additional families from diverse ethnic backgrounds have similar syndromes that have been linked to the chromosome 16 infantile convulsions and choreoathetosis syndrome region. Moreover, in one large pedigree with paroxysmal kinesiogenic dyskinesias only, the syndrome has also been linked to the same genomic area. Families with pure benign familial infantile seizures may be linked to chromosome 16 as well. In this study, we present a series of 19 families and 24 otherwise healthy infants with benign familial infantile seizures. Two of these families include members affected with benign familial infantile seizures and paroxysmal choreoathetosis. We included patients with normal neurologic examinations, who started having simple partial seizures, complex partial seizures, or apparently generalized seizures without recognized etiology between 2 months and 2 years of age. Neurologic studies were normal, but in all patients, there was a history of similar seizures and age at onset in either the father or the mother. Twenty-four patients (14 girls and 10 boys) were evaluated at our hospital between February 1990 and February 2001. Age at onset, sex, family history of epilepsy and/or paroxysmal dyskinesias, neurologic examination, semiology, distribution, and frequency and duration of seizures were evaluated. Electroencephalographic (EEG) and neuroradiologic studies were also performed. Seizures began between 3 and 22 months of life, with a median age of 5 1/2 months. Nine patients (37.5%) had only apparently generalized seizures, 5 patients (20.8%) had only partial seizures, and 10 patients had both partial and apparently generalized seizures (41.6%). Seizures were invariably brief, occurred during the waking state (100%), and presented mainly in clusters in 12 patients (50%). Interictal EEG was normal in 23 patients (95.8%). Sixteen patients (66.6%) had a confirmed history of convulsions in family members other than parents. Twenty-two patients became seizure free after 30 months of life. Two brothers in the same family had brief paroxysmal episodes of choreoathetosis in the hemibody triggered by stress while awake at 15 and 17 years old, respectively. One of them had paroxysmal choreoathetosis only, and the other was associated with benign familial infantile seizures. One father had brief spontaneous episodes of paroxysmal choreoathetosis when awake at age 18 years. All of them had a good response to antiepilepsy drugs, and neurologic examination and EEG and neuroradiologic studies were normal. Benign familial infantile seizure is a genetic epilepsy syndrome with autosomal dominant inheritance. It may be associated with paroxysmal choreoathetosis (infantile convulsions and choreoathetosis syndrome), which has been linked to the chromosome 16 infantile convulsions and choreoathetosis syndrome region. Patients in families with infantile convulsions and choreoathetosis syndrome could display either benign familial infantile seizures or paroxysmal choreoathetosis or both. It is likely that the disease in families with pure benign familial infantile seizures may be linked to the infantile convulsions and choreoathetosis region as well. We cannot exclude the possibility that the youngest patients may develop choreoathetosis or other dyskinesias later in life.     

Efficacy of the ketogenic diet: Which epilepsies respond?

We report the efficacy of the ketogenic diet in refractory epilepsies focusing on outcomes with regard to epilepsy syndromes and etiology in children and adults with refractory epilepsy. Sixty-four consecutive children and four adults were prospectively enrolled from 2002 to 2009; seven were excluded from analysis. The classical ketogenic diet was initiated on an inpatient basis with dietary ratios ranging from 2:1 to 4:1 fat to carbohydrate and protein. Patients were classified according to syndrome and etiology using the 1989 and more recent 2010 International League Against Epilepsy (ILAE) classification systems. Responders were defined as >50% reduction in seizure frequency compared to baseline. Syndromes included symptomatic generalized (52), genetic (idiopathic) generalized (7), and focal epilepsies (2) and etiologies included structural (24), genetic (18), and unknown (19). Twenty-nine (48%) of 61 patients were responders at 3 months. Two children became seizure-free: one with focal epilepsy of unknown etiology and another with refractory childhood absence epilepsy. Responsive syndromes included migrating partial epilepsy of infancy, childhood absence epilepsy, focal epilepsy, epilepsy with myoclonic-atonic seizures, and Dravet syndrome. Children with lissencephaly and hypoxic ischemic encephalopathy had surprisingly good responses. The ketogenic diet is an effective treatment for children and adults with refractory epilepsy. The response is predicted by type of epilepsy syndrome. Accurate characterization of the electroclinical syndrome is an important factor in decisions about timing of initiation of the ketogenic diet.