Spanish Families with Cerebral CavernousAngioma Do Not Bear 742C3T HispanicAmerican Mutation of the KRIT1 GeneMiguel Lucas, MD,* Francisca Solano, BS,*María D. Zayas, PhD,* Jose M. Garcıa-Moreno, MD,†Miguel A. Gamero, MD,† Alzenira F. Costa, PhD,*and Guillermo Izquierdo, MD†

Truncating mutations in CCM1, encoding KRIT1, cause he-reditary cavernous angiomas,1 and 65% of non-HispanicAmerican families have been linked to CCM1 with nofounder effects.2 The common mutation, 742C3T transi-tion, recently described in 16 of 21 families,3 supports thepreviously described strong founder effect in Hispanic Amer-icans.4 We have recently reported on Spanish families withcerebral cavernous angioma linked to CCM1 locus, which donot match the Hispanic American CCM1 haplotype,5 al-though 2 of 9 families partially shared the Hispanic Ameri-can haplotype of CCM1. We have studied the possibility ofan ancestor chromosome in a cohort of Spanish cavernousangioma families by a single screening test that detects the742C3T transition in exon VI of KRIT1 gene.

The exon VI was analyzed in the probands of 39 nuclearSpanish families. A fragment of KRIT1 gene containing exonVI was PCR amplified with primers forward (59 TTGTTAG-ATTGTGATGTA) and reverse (59 AACATAATAAAAACT-TTC). Aliquots of the amplified DNA were initially screenedby analysis of single-strand conformational polymorphism(SSCP). Polymerase chain reaction (PCR) products were elec-trophoresed in 10% acrylamide in the absence and in the pres-ence of 10% glycerol as recently described.1

We analyzed the 742C3T transition by the formation ofthe sequence 59 TTAA that is the target for the restrictionendonuclease MseI. The product of the PCR includes exonVI and an intronic MseI restriction site that served as inter-nal control of the endonuclease digestion. The undigestedPCR product consists of a 238-bp fragment, which, follow-ing the digestion with MseI, is divided in two fragments of210 and 28 bp (Fig). The presence of the 742C3T transi-tion should split the PCR product in three fragments of 141,69, and 28 bp.

DNA sequencing was carried out with 59 32P-labeledprimers and terminal dideoxynucleotides ( fmol kit of Pro-mega, Lyon, France). The forward primer was 59CGAATATA CAGAATGGATG, and the above-describedoligonucleotide was the reverse primer.

The results showed that only 1 of 39 probands gave asignificant SSCP, but both the sequencing of the exon VIand the lack of the MseI restriction site demonstrated thatthe 742C3T transition was not present in the affectedchromosome. Given the great diversity of mutations recentlydescribed in KRIT1 gene,1,3 the finding in the Spanish pop-ulation of a single chromosome with the 742C3T transi-tion could support the existence of an ancestor chromosome.We failed to find this mutation in 39 unrelated patients,however, further supporting that the strong founder effectdescribed in Mexican American families4 with cerebral cav-ernous angioma are specific for this population.

Supported by grant 99/0407 from Fondo de InvestigacionesSanitarias.

Servicios de *Biologıa Molecular and †Neurologıa, HospitalUniversitario Virgen Macarena, Sevilla, Spain

References1. Laberge-le Couteulx S, Jung HH, Labauge P, et al. Truncating

mutations in CCM1, encoding KRIT1, cause hereditary cavern-ous angiomas. Nat Genet 1999;23:189–193

2. Laberge S, Labauge P, Marechal E, et al. Genetic heterogeneityand absence of founder effect in a series of 36 non Hispano-American cerebral cavernoma families. Eur J Hum Genet 1999;7:499–504

3. Sahoo T, Johnson EW, Thomas JW, et al. Mutations in the geneencoding KRIT1, a Krev-1/rap1a binding protein, cause cerebralcavernous malformations (CCM1). Hum Mol Genet 1999;8:2325–2333

4. Gunel M, Awad IA, Finberg K, et al. A founder mutation as acause of cerebral cavernous malformation in Hispanic Americans.N Engl J Med 1996;334:946–951

5. Jung HH, Labauge P, Laberge S, et al. Spanish families withcavernous angioma do not share the Hispano-American CCM1haplotype. J Neurol Neurosurg Psychol 1999;67:551–552

Rapid Diagnosis of Peroxisome BiogenesisDisorders through Immunofluorescence Stainingof Buccal SmearsZhongyi Zhang, MD, Yasuyuki Suzuki, MD,Nobuyuki Shimozawa, MD, and Naomi Kondo, MD

Peroxisome biogenesis disorders (PBDs), including Zellwegersyndrome and neonatal adrenoleukodystrophy, are fatal au-tosomal recessive diseases with severe brain dysfunction anddefective peroxisome biogenesis.1 Diagnosis is usually madeby findings such as an accumulation of very-long-chain fattyacids, deficiency of plasmalogen, and absence of peroxisomesin hepatocytes or cultured fibroblasts. However, liver biopsyis an invasive method, and culture of skin fibroblasts is time-consuming. Buccal smear analysis has been used as noninva-sive materials for sex determination and DNA analysis of ge-netic diseases.2 We reported the detection of peroxisomes inbuccal smears.3 Here we report the usefulness of immuno-

Fig. Analysis of the 742C3T transition in exon VI. MseIdigested and undigested (U) polymerase chain reaction frag-ments of exon VI from different patients were separated underdenaturing conditions in urea-containing 6% acrylamide gel.The image was obtained by autoradiography of the gel. The28-bp fragment, located far below the indicated bands, wasomitted in the photograph.

LETTERS

836 Copyright © 2000 by the American Neurological Association

fluorescence staining of buccal smears for the rapid diagnosisof PBD.

Buccal smears were obtained from healthy control subjectsusing a metal tongue depressor and were variously processedas follows: (1) directly spread onto fluorescence-free coverslips; (2) directly spread onto cover slips and stored at roomtemperature for 48 hours; (3) immediately washed with sa-line, centrifuged, spread onto cover slips; or (4) stored in

saline at 4°C for 48 hours then centrifuged and spread ontocover slips. These cells were fixed with 4% paraformaldehydeand stained by indirect immunofluorescence staining usingantibodies against catalase or acyl-CoA oxidase and FITC-conjugated goat F(ab9)2 anti-rabbit IgG (TAGO, Burlin-game, CA).4

Peroxisomes were clearly and abundantly visualized witheither anti-catalase or anti-acyl-CoA oxidase when the smearswere processed immediately after sampling (Fig, A and C).Peroxisomes were hardly detectable when the cover slips werestored at room temperature for 48 hours (see Fig, B). How-ever, they were detectable when the smear was stored in sa-line at 4°C for 48 hours (see Fig, D). In 2 patients who weresuspected of having Zellweger syndrome, buccal smears insaline were transported at 4°C and stained with anti-catalase24 hours after sampling. Peroxisomes were absent in the firstcase (see Fig, E, left). This patient was later confirmed tohave Zellweger syndrome. The second patient with detect-able peroxisomes was diagnosed as having another congenitalanomaly (see Fig, E, right).

Buccal smears are potential materials for the screening ornoninvasive rapid diagnosis of PBD. Storage of buccalsmears in saline at 4°C for 48 hours conserves the structureof peroxisomes and makes transport of samples possible. Theresults can be obtained within several hours. This methodwill be applicable to the screening of other metabolic neuro-logical diseases.

Department of Pediatrics, Gifu University School of Medicine,Gifu, Japan

References1. Lazarow PB, Moser HW. Disorders of peroxisome biogenesis.

In: Scriber CR, Beaudet AL, Sly WS, Valle D, eds. The meta-bolic and molecular bases of inherited disease. New York:McGraw-Hill, 1995:2287–2324

2. Lench N, Stanier P, Williamson R. Simple non-invasivemethod to obtain DNA for gene analysis. Lancet 1988;1:1356–1358

3. Suzuki Y, Zhang Z, Shimozawa N, et al. Use of buccal smearsfor rapid detection of peroxisomes. Eur J Pediatr 1997;156:250

4. Suzuki Y, Yamaguchi S, Orii T, et al. Nonspecific lipid transferprotein (sterol carrier protein-2) defective in patients with defi-cient peroxisomes. Cell Struct Funct 1990;15:301–308

Topiramate and Essential TremorNestor Galvez-Jimenez, MD, and Melanie Hargreave, RN

Topiramate is a recently approved antiepileptic medicationwith GABAergic, Na1 channel blockade and carbonic anhy-drase inhibition indicated for the treatment of adult partial-onset seizures with and without secondary generalization.Metazolamide, a carbonic anhydrase inhibitor (CAI), hasbeen shown to be useful as add-on therapy or, rarely, asmonotherapy for essential tremor.1 Clonazepam and gaba-pentin,2 both GABAergic medications, may be of some ben-efit in essential tremor. The combined GABAergic and CAIpharmacological properties of topiramate make it suitable forthe treatment of essential tremor.

Nine patients, 4 men and 5 women with an average age of73 years (range, 48–85 years) and disease duration of 18.6

Fig. (A) Buccal smear sample was spread directly onto coverslips. (B) Smear was directly spread onto cover slips and storedat room temperature for 48 hours. (C) Smear was immedi-ately washed with saline, centrifuged, and spread onto coverslips. (D) Smear was stored in saline at 4°C for 48 hoursthen centrifuged and spread onto cover slips. Left side isstained with anti-catalase IgG, and right side is stained withanti-acyl-CoA oxidase IgG. (E, left) Zellweger syndrome;(right) another congenital anomaly. Bar 5 10 mm.

Annals of Neurology Vol 47 No 6 June 2000 837

years (range, 7–40 years), with disabling essential tremorwho failed maximal standard therapy received topiramate.Three patients had associated head tremor and two also hadvoice tremor. In all patients, medications were left un-changed during titration and follow-up period. All patientswere routinely assessed at each visit with the Archimedes Spi-ral (AS) scale,3 the activities of daily living (ADL) self-questionnaire,3 and a visual analogue scale in which theyrated themselves at each visit before and after treatment (Ta-ble). Zero indicated no change, below and above zero indi-cated worse or better control of tremor and disability, respec-tively. In the AS scale (0–10 points), a maximum point of10 indicated complete inability to draw the spiral. Topira-mate was slowly introduced according to the manufacturer’srecommendations, beginning with 50 mg/day for a maxi-mum daily dose of 400 mg. The average dose of topiramatewas 144 mg/day (range, 100–200 mg/day). The total averagescores before and after treatment were 7.7 and 4.8, respec-tively, for the AS scale and 33.8 and 22.0, respectively, forthe ADL scale, resulting in an improvement difference of60% and 53%, respectively. Eight patients rated themselves(visual analogue scale) as better and with less disability aftertherapy. One patient reported increased diuresis while receiv-ing topiramate. The most common side effects were fatigueand paresthesias. Two patients discontinued topiramate ther-apy because of fatigue (Patients 6 and 9; see table) despiteimprovements in tremor. No improvement was seen in heador voice tremors.

Although this is an open-label clinical observation, our ex-perience suggests that topiramate may be useful for the man-agement of essential tremor, especially in patients partiallyresponsive to other established forms of treatment. A pro-spective placebo-controlled trial in drug-naive patients or as

add-on therapy is warranted to assess further the therapeuticimplications of topiramate therapy.

Movement Disorders Program, Department of Neurology,Cleveland Clinic Florida, Fort Lauderdale, FL

References1. Wiener W, Lang AE. Movement disorders: a comprehensive sur-

vey. Mount Kisko, NY: Futura, 19892. Gironell A, Kulisensky J, Barbanoj M, et al. A randomized

placebo-controlled comparative trial of gabapentin and propran-olol in essential tremor. Arch Neurol 1999;56:475–480

3. Bain PG, Findley LJ. Assessing tremor severity. Standards inneurology series. London, Smith-Gordon, 1993

Magnetic Resonance Spectroscopy of EpisodicAtaxia Type 2 and MigrainePasquale Montagna, MD,* Pietro Cortelli, MD,*Raffaele Lodi, MD,† and Bruno Barbiroli, MD†

We wish to draw attention to the similarities between themagnetic resonance spectroscopic (MRS) findings in episodicataxia type 2 (EA2) reported by Sappey-Marinier and col-leagues1 and our own and others’ results obtained in mi-graine patients. Indeed, in our series of 55 patients with mi-graine stroke or migraine with prolonged aura, aura, orwithout aura, 31P-MRS showed a reduced PCr/Pi ratio inoccipital brain regions during the interictal period. Other de-rived indexes of mitochondrial functionality, the adenosinediphosphate, the phosphorylation potential, and the relativerate of energy metabolism, were also abnormal.2 Migraineursalso have increased interictal cerebral lactate on 1H-MRS.3

The MRS changes are thus similar in EA2 and migraine. We

Table. Demographics of Essential Tremor Patients

PatientNo.

Age (yr)/Sex

DiseaseDuration(yr)

TPMX(mg/day) Prior Therapy

Follow-Up(mo) Side Effects

ASS Score,b/a

ADLQ Score,b/a

VASScore

1 78/M 22 100 ppnl, clnz,tznd, mzl

6 Drowsiness, fatigue 8/4 59/28 2

2 85/F 18 100 ppnl, clnz, mzl 12 Fatigue, paresthesia 8/5 25/22 23 48/M 15 250 ppnl, mzl 6 Fatigue 10/6 37/26 24 74/F 40 150 ppnl, mzl,

clnz, BTXa10 Fatigue, paresthesia 9/5 52/18 2

5 83/F 18 100 ppnl, mzl,amtd, BTXa

11 Dry mouth, pares-thesia, somnolence

6/4 24/14 1

6 80/F 7 100 ppnl, mzl,amtd, clnz

2 Fatigue, paresthesia 8/5 27/18 21

7 73/M 20 150 ppnl, mzl,clnz, tznd

12 None 8/5 25/21 1

8 68/M 8 200 ppnl, mzl,clnz, amtd,tznd

3 None 7/5 28/20 2

9 68/F 20 150 ppnl, mzl, clnz 3 Diuresis, fatigue,nausea

6/5 28/27 1

ppnl 5 propranolol; clnz 5 clonazepam; tznd 5 tizanidine; mzl 5 myzoline; amtd 5 amantadine; BTXa 5 botulinum toxin type A; ASS 5Archimedes Spiral scale; b/a 5 before/after; ADLQ 5 activities of daily living self-questionnaire; VAS 5 visual analogue scale at last follow-up.

838 Annals of Neurology Vol 47 No 6 June 2000

took our findings to indicate that migraine patients have anunderlying disorder of mitochondrial oxidative metabolism,which may make their brains more susceptible to metabolicstress and metabolic exhaustion, and which may lead to aninability to maintain physiological ionic traffic and restingand action potentials during the attacks. Altered ion levels inthe form of low [Mg21]i and [H1]i are indeed detected by31P-MRS in migraine.4 When we consider that EA2 is a dis-ease allelic to familial hemiplegic migraine (FHM), caused bymutations in the same CACNA1A gene, and that 31P-MRSshowed comparable abnormalities also in a family withFHM, admittedly without genetic verification,5 the conclu-sion is that EA2 and migraine share similar abnormalities ofoxidative mitochondrial metabolism. If migraine is indeed achannelopathy, it is conceivable that enhanced Ca21 entrythrough faulty ion channels and Ca21 overload in the mito-chondria may impair energy production and ion levels, lead-ing to a vicious circle of metabolic and ionic instability. Theother way around may also, however, prove true—that is, aprimary impairment of mitochondrial energy function lead-ing to instability in maintaining regular neuronal ionic levelsand subsequent depolarization, Ca21 entry, and possiblyspreading depression. We could thus explain the migrainousfeatures of some primarily mitochondrial DNA diseases, suchas MELAS, in which ion channel abnormalities are notreadily apparent.

*Institute of Clinical Neurology and †Department of ClinicalMedicine and Applied Biotechnology “D. Campanacci,”University of Bologna, Bologna, Italy

References1. Sappey-Marinier D, Vighetto A, Peyron R, et al. Phosphorus and

proton magnetic resonance spectroscopy in episodic ataxia type2. Ann Neurol 1999;46:256–259

2. Montagna P, Cortelli P, Barbiroli B. Magnetic resonance spec-troscopy studies in migraine. Cephalalgia 1994;14:184–193

3. Watanabe H, Kuwabara T, Ohkubo M, et al. Elevation of cere-bral lactate detected by localized 1H-magnetic resonance spec-troscopy in migraine during the interictal period. Neurology1996;47:1093–1095

4. Lodi R, Montagna P, Soriani S, et al. Deficit of brain and skel-etal muscle bioenergetics and low brain magnesium in juvenilemigraine: an in vivo 31P magnetic resonance spectroscopy inter-ictal study. Pediatr Res 1997;42:866–871

5. Uncini A, Lodi R, DiMuzio A, et al. Abnormal brain and muscleenergy metabolism shown by 31P-MRS in familial hemiplegicmigraine. J Neurol Sci 1995;129:214–222

ReplyDominique Sappey-Marinier, PhD,Emmanuel Broussolle, MD, and Alain Vighetto, MD

We thank Dr Montagna and his colleagues for their valuablecomments. Indeed, we were aware of similar magnetic reso-nance spectroscopic (MRS) findings between episodic ataxiatype 2 (EA2) patients and migraine patients as reported bythe group of Montagna1–3 and others.4,5 Because of spacelimitation, however, we were unable to mention these resultsin our article.6 Montagna and associates1 showed by 31PMRS a reduced PCr content in occipital brain of patientswith migraine with and without aura during the interictal

period, and Uncini and co-workers3 reported similar meta-bolic alterations in familial hemiplegic migraine patients.Watanabe and colleagues4 demonstrated an increase in lac-tate peak in the occipital brain of 5 patients who had expe-rienced a migraine attack within the previous 2 months,whereas a sixth patient who had no attack in the previous 4years showed no lactate peak. Ramadan and associates5

found a low brain magnesium concentration in migraine.These metabolic changes, detected by 1H or 31P MRS, arecomparable with our findings in EA2 patients.6 Consideringalso the observations of mutations on the same CACNA1Agene in EA2 and familial hemiplegic migraine,7 these resultsraise the question of the underlying pathophysiologicalmechanisms. Whether a possible Ca21 overload due to chan-nelopathy may primarily lead to ionic instability or to mito-chondrial energy dysfunction remains unclear. Althoughseveral studies support the “mitochondrial hypothesis of mi-graine,”8,9 the observation of a pH increase in EA2 patientsin contrast to migraine patients may strengthen the “ionicinstability hypothesis of EA2.” However, further investiga-tions are needed to determine the relationship between theobserved gene mutations and metabolism alterations in orderto better understand the mechanisms of these brain diseases.

Hopital Neurologique Pierre Wertheimer, Universite ClaudeBernard Lyon 1, Lyon, France

References1. Montagna P, Cortelli P, Barbiroli B. Magnetic resonance spec-

troscopy studies in migraine. Cephalgia 1994;14:184–1932. Lodi R, Montagna P, Soriani S, et al. Deficit of brain and skel-

etal muscle bioenergetics and low brain magnesium in juvenilemigraine: an in vivo 31P magnetic resonance spectroscopy inter-ictal study. Pediatr Res 1997;42:866–871

3. Uncini A, Lodi R, DiMuzio A, et al. Abnormal brain and muscleenergy metabolism shown by 31P MRS in familial hemiplegicmigraine. J Neurol Sci 1995;129:214–222

4. Watanabe H, Kawabara T, Ohkubo M, et al. Elevation of cere-bral lactate detected by localized 1H magnetic resonance spec-troscopy in migraine during the interictal period. Neurology1996;47:1093–1095

5. Ramadan NM, Halvorson H, Vande-Linde A, et al. Low brainmagnesium in migraine. Headache 1989;29:416–419

6. Sappey-Marinier D, Vighetto A, Peyron R, et al. Phosphorus andproton magnetic resonance spectroscopy in episodic ataxia type2. Ann Neurol 1999;46:256–259

7. Ophoff RA, Tewerwindt GM, Vergouwe MN, et al. Familialhemiplegic migraine and episodic ataxia type-2 are caused bymutations in the Ca21 channel gene CACNL1A4. Cell 1996;87:543–552

8. Lanteri-Minet M, Desnuelle C. Migraine et dysfonction mito-chondriale. Rev Neurol 1996;152:234–238

9. Bresolin N, Martinelli P, Barbiroli B, et al. Muscle mitochon-drial DNA deletion and 31P NMR spectroscopy alterations in amigraine patient. J Neurol Sci 1991;104:182–189

Clinical Heterogeneity in Pedigrees with2q-Linked Febrile SeizuresBruno Moulard, MD, PhD,* Denys Chaigne, MD,†and Alain Malafosse, MD, PhD*

We read with interest Peiffer and colleagues’ article1 report-ing on a Utah family with febrile seizures (FSs) linked to

Annals of Neurology Vol 47 No 6 June 2000 839

chromosome 2q23–24. Indeed, we,2 and others,3,4 have re-cently reported linkage between a specific subgroup of famil-ial FSs, known as generalized epilepsy and febrile seizures“plus” syndrome (GEFS1), and markers in the same regionof chromosome 2q. Peiffer and co-workers1 conclude thatthe trait segregating in the Utah family is different fromGEFS1 and consider it as familial FS, naming the locusFEB3. The regions containing both the 2q-linked GEFS1

gene and that responsible for FSs in the Utah family arelarge. These phenotypes may therefore be determined by twodifferent genes. However, the hypothesis that the Utah fam-ily is the fourth GEFS1 pedigree linked to chromosome 2qmust also be considered.

GEFS1 is a recently identified and clinically very hetero-geneous familial autosomal dominant epilepsy syndrome.5

Patients present with either isolated FSs or with FSs “plus”(FS1). Thus, they have FSs after the age of 6 years with orwithout subsequent development or co-occurrence of afebrileseizures, usually generalized tonic/clonic seizures (GTCSs).3,5

In the first GEFS1 pedigree to be reported,5 originatingfrom the Victoria region of Australia and linked to chromo-some 2q,3 the most common phenotype was FS1. The au-thors further characterized the GEFS1 phenotype using 9Australian families.3 They observed that half the affected rel-atives presented with isolated FSs. In both of the other re-ported GEFS1 pedigrees linked to chromosome 2q,2,4 thephenotype was different.

In our GEFS1 family,2 more than half the affected rela-tives had isolated FSs, whereas in the other French GEFS1

family,4 only 1 relative had isolated FSs. Another strikingdifference was the severity of epilepsy. In our family,2 noaffected relatives had epilepsy after their teenage years, exceptone, whose last seizure occurred at the age of 20 years. In theother French family,4 in contrast, epilepsy continued intoadulthood in most of the affected subjects. In the Utah fam-ily, half the affected relatives (9/17, 53%) had isolated FSsand the others (8/17, 47%) had FS until the age of 5 to 6years and afebrile seizures (GTCS or partial seizures) there-after, with or without a seizure-free interval. Since FSs wereabsent after the age of 6 years, the authors did not considerthis phenotype to be GEFS1. It should be stressed, however,that at least 2 affected relatives in the 2 French GEFS1 fam-ilies2,4 did not have FSs after the age of 6 years, but afebrileseizures, as reported in the Utah family. It is also noteworthythat patients with sporadic or familial FSs can present sub-sequent afebrile seizures. However, such cases represent aslight percentage of FS cases, whereas in the Utah familyalmost half the affected cases had subsequent afebrile sei-zures, as observed in the other GEFS1 families. Finally, theUtah family and our family2 are very similar: mostly isolatedFSs (53% and 55%, respectively) and benign progression ofthe syndrome without seizures after the adolescence.

In conclusion, with the reported Utah family, the suscep-tibility locus for FS on chromosome 2q may be considered tobe a major susceptibility locus for the very heterogeneous ep-ilepsy syndromes with FSs. It is also possible that some fam-ilies, especially nuclear families, with apparently isolated id-iopathic generalized epilepsy, such as apparently familialisolated FS cases, are actually GEFS1 pedigrees.

*Division de Neuropsychiatrie, Unite de Biochimie Genetique,Hopital Belle Idee, Hopitaux Universitaires de Geneve,Chene-Bourg, Switzerland, and †Clinique Sainte-Odile,Strasbourg, France

References1. Peiffer A, Thompson J, Charlier C, et al. A locus for febrile

seizures (FEB3) maps to chromosome 2q23–24. Ann Neurol1999;46:671–678

2. Moulard B, Guipponi M, Chaigne D, et al. Identification of anew locus for generalized epilepsy with febrile seizures plus(GEFS1) on chromosome 2q24–q33. Am J Hum Genet 1999;65:1396–1400

3. Singh R, Scheffer IE, Crossland K, Berkovic SF. Generalized ep-ilepsy with febrile seizures plus: a common childhood-onset ge-netic epilepsy syndrome. Ann Neurol 1999;45:75–81

4. Baulac S, Gourfinkel-An I, Picard F, et al: A second locus forfamilial generalized epilepsy with febrile seizures plus (GEFS1)maps to chromosome 2q21–q33. Am J Hum Genet 1999;65:1078–1085

5. Scheffer IE, Berkovic SF. Generalized epilepsy with febrile sei-zures plus: a genetic disorder with heterogeneous clinical pheno-types. Brain 1997;120:479–490

Locus for Febrile SeizuresIngrid E. Scheffer, MBBS, PhD,*Robyn H. Wallace, PhD,† John C. Mulley, PhD,†and Samuel F. Berkovic, MD, FRACP*

Peiffer and colleagues1 have reported on a large family withfebrile and afebrile seizures that map to chromosome 2q.They argue that this represents a third locus for the commonform of febrile seizures rather than a locus for the newlydescribed syndrome of generalized epilepsy with febrile sei-zures “plus” (GEFS1). We disagree. On clinical grounds,their family has GEFS1, and this is supported by recent mo-lecular genetic data.

GEFS1 is a genetic epilepsy syndrome characterized byheterogeneous epilepsy phenotypes. The phenotypes mostcommonly seen are typical febrile seizures (FSs) and FSs“plus” (FS1), with more severe phenotypes such as myoclonic-astatic epilepsy occurring less frequently.2–4 Individuals withthe FS1 phenotype have febrile seizures that extend outsidethe classical age-related definition of FSs, occurring between 3months and 6 years, or they have additional afebrile seizures.The critical feature is the continuity of generalized seizuresfrom early to mid childhood, not the presence or absence offever. Indeed, families often do not recognize the absence of afever, interpreting all attacks as “febrile seizures.” This may inpart relate to afebrile seizures carrying the stigma of epilepsy.

Peiffer and associates1 argue that their family “differ[s]substantially” from GEFS1 because none of their cases hadseizures with fever beyond the age of 6 years, whereas theybelieve that most patients in our original family did.1 Theymisunderstood our initial report.2 In our original family, theparents were asked if more or less than 50% of seizures oc-curred with fever. In some children with FS1, seizures withfever did occur beyond age 6 years; in others, all seizuresbeyond age 6 years were afebrile as in Peiffer’s kindred. Con-tinuity into mid childhood was present in Peiffer’s fam-ily—in 6 of 8 individuals afebrile seizures followed within 2years of the last febrile seizure.

840 Annals of Neurology Vol 47 No 6 June 2000

Peiffer and co-workers also did not refer to our later studyof 9 families,3 in which we pointed out that the most com-mon phenotype in many GEFS1 families is in fact febrileseizures that do cease before age 6 years. Moreover, no lessthan 10 of 18 older individuals from their pedigree had fe-brile seizures up to the age of 5 or 6 years, and, as indicatedin their review, 90% of febrile seizures occur within the first3 years of life.1 The cutoff for the typical febrile seizure syn-drome at 5 or 6 years is arbitrary. Although none of theircases were known to have febrile seizures beyond 6 years, theage distribution of febrile seizures in their family was atypicalfor the common febrile seizure syndrome but was perfectlyconsistent with GEFS1.

Although partial seizures were not included in our originalreport delineating GEFS1,2 these have subsequently been re-ported,3,7 indicating that partial seizures can occur in theGEFS1 spectrum. Thus, their individuals with later partialseizures1 would also fit into this genetic epilepsy syndrome.

In addition to these clinical arguments, there are at least 5other GEFS1 families that map to the same region on 2q.These include our original Australian family in whomGEFS1 was described,4–6 2 French families with typicalGEFS1,7,8 and 2 families from the Middle East (unpub-lished data). While it is possible that there are a number of“febrile seizure genes” in this region, a more parsimoniousexplanation that accords with our clinical analysis is thatthere is a single gene in that region that is responsible for asizeable proportion of GEFS1 families. Unless there is mo-lecular evidence of more than one gene locus for seizureswith fever in the chromosome 2q region, this locus should bedesignated GEFS2 rather than FEB3. As pointed out by anumber of groups,1,6–8 the cluster of sodium channel sub-units at the GEFS2 locus are excellent candidate genes, fol-lowing the finding of a mutation in the neuronal sodiumchannel subunit SCN1B at GEFS1 on chromosome 19.9

*Department of Medicine (Neurology), University ofMelbourne, Austin and Repatriation Medical Centre,Heidelberg, Victoria, Australia; and †Department ofCytogenetics and Molecular Genetics, Centre for MedicalGenetics, Women’s and Children’s Hospital, Adelaide,South Australia

References1. Peiffer A, Thompson J, Charlier C, et al. A locus for febrile

seizures (FEB3) maps to chromosome 2q23–24. Ann Neurol1999;46:671–678

2. Scheffer IE, Berkovic SF. Generalized epilepsy with febrile sei-zures plus: a genetic disorder with heterogeneous clinical pheno-types. Brain 1997;120:479–490

3. Singh R, Scheffer IE, Crossland K, Berkovic SF. Generalized ep-ilepsy with febrile seizures plus (GEFS1): a common, childhoodonset, genetic epilepsy syndrome. Ann Neurol 1999;45:75–81

4. Berkovic SF, Scheffer IE. Febrile seizures: genetics and relation-ship to other epilepsy syndromes. Curr Opin Neurol 1998;11:129–134

5. Lopes-Cendes I, Scheffer IE, Berkovic SF, et al. Mapping a locusfor idiopathic generalized epilepsy in a large multiplex family.Epilepsia 1996;35:127 (Abstract)

6. Lopes-Cendes I, Scheffer IE, Berkovic SF, et al. A new locus forgeneralized epilepsy with febrile seizures plus maps to chromo-some 2. Am J Hum Genet 2000 (in press)

7. Baulac S, Gourfinkel-An I, Picard F, et al. A second locus forfamilial generalized epilepsy with febrile seizures plus maps tochromosome 2q21–q33. Am J Hum Genet 1999;65:1078–1085

8. Moulard B, Guipponi M, Chaigne D, et al. Identification of anew locus for generalized epilepsy with febrile seizures plus(GEFS1) on chromosome 2q24–q33. Am J Hum Genet 1999;65:1396–1400

9. Wallace RH, Wang DW, Singh R, et al. Febrile seizures andgeneralized epilepsy associated with a mutation in the Na1-channel b1 subunit gene SCN1B. Nat Genet 1998;19:366–370

CorrectionThe primer sequence given in the second paragraph of theMethods section of the article by Pulkes et al in the Decem-ber issue (Ann Nuerol 1999;46:916–919) was incorrect. Thecorrect sequence is TGC GGT TTC GAT GAT GAT AT.

Annals of Neurology Vol 47 No 6 June 2000 841