Epilepsia. 33(Suppl. 4):S26-S3 I , 1992 Raven Press, Ltd., New York 0 International League Against Epilepsy
Treatment of Infantile Spasms: Medical or Surgical? *tW. Donald Shields, *?D. Alan Shewmon, *?$.Harry T. Chugani, and §Warwick J. Peacock Departmenls of +Neurology, f Pediatrics, $Radiology, and §Neurosurgery, University of CaliJornia, Los Angeles, School of Medicine, Los Angeles, California, U.S.A.
Summary: Although infantile spasms were initially described in 1 84 I , remarkably little progress has been made in understanding the pathophysiology of this “peculiar form of infantile convulsions.” Consequently, our ability to treat infantile spasms is limited. Infantile spasms are classified as a “generalized” seizure disorder in the international classification system, which suggests that the underlying brain abnormality causing the seizures also must be diffuse or generalized. As the classification suggests, there are many diffise, or multifocal, brain disorders related to infantile spasms, e.g, inborn errors of metabolism, hypoxic-ischemic brain injury, and developmental brain defects such as tuberous sclerosis or Aicardi’s syndrome. On the other hand, infantile spasms have been reported in which a localized brain abnormality was present, e.g., tumor, stroke, and trauma. On rare occa-
In 1841 a British physician, W. J. West, wrote the first description of infantile spasms in a letter to the editor of Lancet titled “On a Peculiar Form of Infantile Convulsions.” He described in a young child (his own son) the seizures that we now recognize as infantile spasms, including their most distressing associated feature-developmental decline (West, 184 1). It is surprising how little has been learned of the basic mechanisms of these seizures in the 150 years since the original description. Perhaps as a consequence of our lack of understanding, few effective medical therapies are available, and the majority of the children still suffer the grave developmental consequences Dr. West described. Epilepsy in some children can be described as “catastrophic” because of the impact it has on their lives. Catastrophic childhood epilepsy is characterized by (a) onset early in life (often by age 1 year); (b) inadequate
sions, removal of a tumor has resulted in cessation of the generalized infantile spasms. This finding suggests that focal cortical abnormalities can cause infantile spasms and that removing the abnormality can stop the seizures. At University of California, Los Angeles, the Pediatric Epilepsy Surgery Program has developed new approaches to the treatment of infantile spasms. The principal underlying concepts are (a) children with medically refractory infantile spasms may have an area of cortical defect (called the zone of cortical abnormality) that causes the seizures and (b) infantile spasms are usually generalized seizures. Thus, the goal of the surgical assessment is not the identification of the focus of seizure onset but rather the identification of the zone of cortical abnormality. Key Words: Infantile spasms-AnticonvulsantsNeurosurgery.
response to antiepileptic drugs (AEDs); and (c) a poor developmental outcome. The most important catastrophic childhood seizure disorder is that of infantile spasms. The seizures of infantile spasms consist of brief bilateral jerks of the trunk and limbs, either flexion, extension, or a combination of both. Hypsarrhythmia, an electroencephalographic (EEG) abnormality defined as generalized, high-voltage, chaotic slowing with very high-voltage multifocal epileptiform discharges, is usually present during the interictal state but its presence is not required for making the diagnosis. The widespread EEG abnormalities combined with the bilateral clinical appearance of the seizures has led to the classification of infantile spasms as a generalized seizure disorder (Commission, 1989). Although neither the physiologic mechanisms of the seizures nor the specific reasons for the developmental consequences are understood, much has been learned about the neurologic diseases associated with infantile spasms. It is important to recognize that infantile spasms are a manifstation of an underlying brain disorder rather than a disease entity unto itself. Thus, recognition that a child has infantile spasms is the beginning of the evaluation rather than the conclusion. That
Address correspondence and reprint requests to Dr.W.Donald Shields at Division of Neurology, Department of Pediatrics, UCLA School of Medicine. 10833 Le Cone Avenue, Los Angeles, CA 90024-1752, U.S.A.
S26
TREATMENT OF INFANTILE SPASMS
diverse brain abnormalities are assOciated with infantile spasms is evidence, at least in some cases, that it is a secondary generalized seizure disorder rather than a primary one. On the basis of the identified or presumed cause of the seizures, patients with infantile spasms are divided into two groups: those with symptomatic and those with cryptogenic seizures. In symptomatic cases, an underlying cause can be identified, and in cryptogenic cases, no predisposing abnormality can be discovered after complete evaluation. Virtually anything that can injure the developing brain has been described as a cause of infantile spasms. As one might expect, many of these causes are diffuse or multifocal (e.g., hypoxic brain damage, diffuse brain malformation, inborn error of metabolism), but others are unifocal brain disturbances (e.g., tumor, stroke, trauma, focal brain malformation). Regarding patients who have a unifocal brain disorder such as tumor or stroke, one can ask an important question: Is the seizure disorder a manifestation of an undetected generalized brain lesion or can a unifocal brain disorder result in a generalized seizure lesion such as infantile spasms? A corollary question logically follows: If the patient’s seizures are medically refractory, can resection of the unifocal brain lesion result in cessation of the infantile spasms? MEDICAL TREATMENT OF INFANTILE SPASMS
Infantile spasms are often difficult to control with medications (Bachman, 1982; Dreihss et al., 1986; Hrachovy and Frost, 1989).One factor setting infantile spasms apart from all other seizure disorders is the superior efficacy of steroid drugs rather than AEDs in treatment (Dreifuss, 1983;Ferry et al., 1986;Holmes, 1987;Engel, 1989).There is disagreement as to whether corticotropin is more effective than prednisone, but most authors recommend 20-40 units of corticotropin per day given intramuscularly as the initial treatment. Many patients respond dramatically to corticotropin, with cessation of seizures after one or two doses, but treatment failures are common. Medications reported to be occasionally effective include pyridoxal phosphate (Blennow and Stark, 1986;Ohtsuka et al., 1987),highdose corticotropin (>lo0U/day) (Riikonen and Donner, 1980;Snead et al., 1983;Snead, 1990), highdose valproate (PaNone et al., 1981; Dyken et al., 1985; Siemes et al., 1988), and nitrazepam, y-vinyl-yaminobutync acid (Chiron et al., 1989;Livingston et al., 1989;Auerswald et al., 1991).Infantile spasms also have been reported to stop spontaneously, so evaluation of a therapeutic regimen that controls only a small percentage of patients is difficult.
S2 7
PROGNOSIS
The prognosis for normal development in children who develop infantile spasms is poor. Overall, 90-95% of children with infantile spasms will be moderately or markedly mentally retarded by age 5 years (Glaze et al., 1988).The prognosis appears to be related primarily to the cause and to seizure control (Riikonen, 1982; Fois et al., 1984). Other factors, such as age of onset of seizures, the presence of seizures other than infantile spasms, and localized EEG, computed tomographic (CT), and magnetic resonance imaging (MRI) abnormalities have also been related to outcome. The most important predictor of outcome is the underlying cause. In about 60-70% of patients, the cause is categorized as symptomatic and in 30% as cryptogenic. The probability of normal outcome for children with symptomatic infantile spasms is only about 5%, whereas it is as high as 38% in those with cryptogenic seizures (Glaze et al., 1988). In either group, inability to control the seizures indicates a poorer prognosis for normal development. Onset of seizures before the child is 4-6 months of age is not a good sign. In one study, seven of nine patients with an onset before 6 months of age were severely handicapped at follow-up, whereas only one of 16 with onset at the age of 6 months or older had such severe handicaps (Seki et al., 1976). In another study, one of 78 children with onset before 4 months of age developed normally, compared with 30 of 195 children with seizure onset at 4 months or older (Lombroso, 1983). The Occurrence of other types of seizures in addition to infantile spasms is also unfavorable. Three studies found that if other types of seizures occurred before or after infantile spasms,15% or fewer patients had a good outcome, whereas if infantile spasms were the only seizure type,more than 30% of the patients had a good outcome (Ohtahara et al., 1980; Matsumoto et al., 1981; Riikonen, 1982). Thus, patients who have medically refractory symptomatic infantile spasms and onset of seizures before the age of 4-6 months have a poor prognosis for normal development. For this group of patients, new therapies are certainly needed. SURGICAL TREATMENT: BACKGROUND
Evidence of localized brain abnormalities has been reported in many patients with infantile spasms. Focal disturbances have been demonstrated on EEG, localized structural defects by CT and MRI,and focal metabolic abnormalities by positron emission tomography (PET).EEG focal disturbance is the most commonly reported abnormality. Such focal abnormalities were EpiIepsia, Vd.33, Supd. 4, I992
W. D. SHIELDS ET AL.
S28
noted in one series in 110 of 149 cases (Riikonen, 1982). Because the EEG of a child with infantile spasms is so abnormal, it is easy to miss a relatively subtle EEG abnormality such as focal slowing or decreased activity. Structural defects detected by CT or MRI are less common but usually much more obvious (Singer et al., 1982). Alvarez et al. (1987) reported three children with porencephaly who developed infantile spasms and proposed that the porencephaly was the cause of the infantile spasms. Cusmai et al. (1988) found CT evidence of focal abnormalities in 17 of 174 cases. In a retrospective review of 37 patients with infantile spasms, we observed definite localized abnormalities on CT or MRI in 17 cases (46%) and EEG evidence of focality in another 6 (16%). Only 38% of the patients had no localizing features (Shields et al., 1992). The percentage of localizing features we observed is probably an overestimate of the true incidence, as some of the patients were referred because they had evidence of a localizable lesion and were thus candidates for surgery. Nevertheless, localized abnormalities may be much more common than previously appreciated. There are a few case reports of children with infantile spasms in whom a localized brain lesion was observed on neuroimaging and whose spasms stopped when the lesion was resected. Infantile spasms ceased after surgical resection of a right temporal astrocytoma in one case (Mimaki et al., 1983), by resection of a choroid plexus papilloma in another (Branch and Dyken, 1979), and by fenestration of porencephalic cysts in four cases (Palm and Korinthenberg, 1988).
SURGICAL TREATMENT: CONCEP'IS The Pediatric Epilepsy Program at University of California, Los Angeles (UCLA) was developed for the special needs of children with seizures. Pediatric epilepsy surgery evolved as part of the program and as a complement to the adult epilepsy surgery program. New concepts and protocols have been developed for addressing the unique challenges presented by children with medically refractory seizures. During the past 30 years, surgery to control medically intractable epilepsy has been directed toward patients with partial seizures (which may generalize secondarily). The essential observation that permits surgery is the identification of the area of seizure onset by surface EEG or by the placement of electrodes in or on the brain. Patients with generalized seizures (either primary or secondarily) have therefore not qualified for f d resective surgery. One result of our program has been the recognition that some patients with generalized seizures such as infantile spasms have a localizable Epilepsia Vd.33, Supd. 4, 1992
cortical abnormality that can be resected, with consequent control of seizures.The ability to perform surgery in these patients has necessitated fundamental changes in the concepts used in identifying the area to be removed. Through the study of a series of patients, we have determined factors related to the underlying disorders that sometimes cause infantile spasms and have developed new protocols that identify the area of abnormal cortex responsible for the seizures. We have called this area the zone of cortical abnormality to distinguish it from the epileptic focus or epileptic zone used to identify surgical candidates among patients with partial seizures. At UCLA, cortical resection has become an accepted method of therapy for medically intractable infantile spasms. The two concepts that govern surgical therapy in children with infantile spasms can be summarized as follows. Concept 1: Because the patients may have generalized or multifocal bihemispheric seizures, the goal of presurgical assessment is not the identification of the focus of onset of seizures but rather the identification of an area of structurally abnormal cortex called the zone of cortical abnormality. Concept 2: Resection of the zone of cortical abnormality in properly selected patients can result in the cessation of generalized seizures, including infantile spasms.
SURGICAL CANDIDATES Children are considered for evaluation for surgical control of seizures if they have infantile spasms that are refractory to corticotropin and/or prednisone and to trials with AEDs. We require that no evidence exists of degenerative, metabolic, or storage diseases, or medical contraindications for surgery. The majority of the patients have evidence of a localized brain abnormality before undergoing formal surgical evaluation. Most patients had partial seizures that antedated the onset of infantile spasms and/or had focal abnormalities on EEGs performed before the development of hypsarrhythmia. We have evaluated several patients who did not have preexisting evidence of localized brain defect and have not been able to identify a zone of cortical abnormality in the majority of such patients.
SURGICAL EVALUATION Evidence of a zone of cortical abnormality is sought by careful neurologic examination, by review of the history of seizures for determining whether partial seizures preceded or followed the infantile spasms, and by examination of present and past EEGs for identifying f d abnormalities. CT and MRI scans also are examined for evidence of a focal structural lesion. If the quality of CT and MRI scans is inadequate, they
TREA TMENT OF INFANTILE SPASMS
are repeated. Children whose results of neuroimaging studies reveal diffuse damage affecting both cerebral hemispheres are not further considered for surgery. The children are then admitted for in-hospital, 24h/day closed-circuit television EEG (CCTVIEEG) to characterizeand quantify seizures. Because the children have very frequent seizures, the duration of CCTV/ EEG is usually much shorter than the typical evaluation for epilepsy surgery. While the child is hospitalized, a PET scan is performed to evaluate the metabolic activity of the brain, particularly identification of focal hypometabolism.
SURGICAL CRlTERIA Children are recommended for surgery if there is a confluence of data from their history and EEG, C W / EEG, CT, MRI, and PET studies indicating the presence of a zone of cortical abnormality (a single lobe, contiguous multilobes, or one hemisphere). No single abnormal result is sufficient for surgery. A surgical decision is based on the following criteria.
EEGs and preoperative CC”/EEG (a) If the child had partial seizures prior to infantile spasms, the area of onset is identified by reviewing the description of the seizures and the past EEGs. (b) If the child has partial seizures at the time of surgical evaluation, the CCTV/EEG is reviewed to identify the focus of onset of partial seizures. (c) If the child has generalized or multifocal seizures at the time of surgical evaluation, the CCTV/EEG is analyzed for the presence of interictal abnormalities associated with a structural defect (focal slowing, a predominant focus of interictal epileptiform discharges, decreased P activity, or unilateral subclinical electrodecremental events). Imaging studies (a) The CT and/or MRI scans are examined for evidence of a zone of cortical abnormality. The abnormalities may be quite subtle (e.g., thickened cortex, narrow white matter, or suIci oriented in a wrong direction). If a zone of cortical abnormality is observed, it must correlate with the EEG focus. However, a completely normal CT and/or MRI result is not a contraindication for surgery. (b) The PET study must demonstrate a focal or hemispheric hypometabolic area (or hypermetabolic if it is an ictal study rather than i n t e n d ) that correlates with the focal EEG abnormality and with the CT/MRI (if abnormal). If the PET is the only study indicating focality, the child is not considered to be a candidate for surgery.
S29
(c) The brain outside the area of possible resection should appear to be relatively normal on CT,MRI, and PET (excluding generalized atrophy due to corticotropin therapy). Neurologic examination If a focal neurologic deficit is observed on examination, it must correlate with the zone of cortical abnormality being considered for resection. Additional neuropbysiologic studies If the child is a likely surgical candidate but the above do not define the zone of cortical abnormality sufficiently, the following studies are performed: (a) somatosensory-evoked potentials, observing for unilateral decreased or absent cortical peak; (b) thiopental sodium test showing a focal area of impaired drug-induced B activity on EEG, (c) intracarotid amobarbital injection on the “bad side,” resulting in suppression of multifiocal independent spikes bilaterally (Shewmon et al., 1990). Assessment of anticipated neurologic deficit The cortical resection cannot create an unacceptable new neurologic deficit. There is, however, no absolute definition of “unacceptable.” A decision is made in consultation with the patient’s parents on the basis of the severity of the seizures, the extent of developmental failure, and the prognosis if the seizures are to remain uncontrolled, compared with the effects of the anticipated deficit. For example, children considered for hemispherectomy are required to manifest hemiparesis prior to surgery. Thus, creation of a new severe hemiparesis is not considered “acceptable.” However, creation of a homonymous hemianopsia is usually considered acceptable.
SURGICAL TREATMENT: RESULTS Since 1986, we have performed surgery on 20 patients who bad a history of infantile spasms. The first four patients had partial seizures at the time of the surgery but had a history of generalized infantile spasms (Shields et al., 1992). This finding suggested that some children with infantile spasms had a localized cortical abnormality associated with the generalized seizures. The generalized seizures were controlled with corticotropin, but catastrophic partial seizures persisted. We questioned whether there were children with medically refractory generalized infantile spasms who had a similar localized abnormality that could be removed, with consequent control of the generalized seizures. In 1987, a patient with tuberous sclerosis presented with medically refractory generalized infantile spasms. An MRI scan demonstrated the typical areas of signal intensity in the cortex, but there was also a cortical glioma in the left parietal lobe that required resection. Epilepsia Vd.33, Suppl. 4. 1992
S30
W.D. SHIELDS ET AL.
The EEG revealed that the area near the glioma had a very active interictal epileptogenic focus. We determined that if this patient’s infantile spasms resolved with resection of the cortical glioma, the essential concept of control of generalized seizures by resection of a localized cortical abnormality would be supported. The glioma was resected, and the patient was seizurefree for 7 months. Following a period of seizure recurrence that was associated with the discontinuation of nitrazepam therapy, he again became seizure-free and has been so for the last I $ years. We subsequently extended the surgery protocol to include patients with medically refractory generalized infantile spasms even without associated focal seizures. Of the next 16 children, 13 had generalized infantile spasms at the time of surgery. The other three had mixed partial seizures and infantile spasms. The results with respect to seizure control have been quite gratifying. Most patients have a large zone of cortical abnormality requiring multilobar resection or hemispherectomy (Shields et al., 1992). Of the initial nine previously reported patients who were treated surgically for medically intractable infantile spasms, seven remain free of seizures 3-6 years after surgery. The surgery is as effective in patients who have generalized seizures as in patients who have partial seizures at the time of surgery (Shields et al., 1990). Several patients had cryptogenic infantile spasms with normal findings on CT and MRI Scans and no medical history suggesting a cause for the infantile spasms. The four reported by Chugani et a!. (1 990) all became seizurefree after surgery. CONCLUSIONS
The obvious answer to the question posed by the title of this article is that both types of treatment, medical or surgical, could be appropriate, depending on the specifics of the patient. Treatment, however, occurs sequentially-first medical and then the question of possible surgery. Without question, the initial treatment of choice is medical. Administration of corticotropin and/or prednisone should begin immediately after the diagnosis of infantile spasms has been established. Children who do not respond to this initial therapy should be given AEDs. Those whose seizures are controlled with medications are not surgical candidates whether or not they have an identifiable zone of cortical abnormality. On the other hand, the prognosis for children for whom these medical therapies fail is quite dismal. Consequently, the risks of surgery become acceptable when medical therapies have failed. Most patients with medically intractable infantile spasms are probably not candidates for surgery. The Epilepsia. Vol. 33. Suppl. 4,1992
incidence of the presence of a zone of cortical abnormality in children with medically refractory infantile spasms has not been established. However, on the basis of our experience, many such patients will benefit from surgery. We offer the following conclusions. (a) Some children with medically refractory infantile spasms have a localized brain defect, which we have called the zone of cortical abnormality. (b) Children with a zone of cortical abnormality may have generalized seizures. (c) Children with a zone of cortical abnormality may have normal CT and MRI scans. (d) The zone of cortical abnormality is often best identified with use of data obtained by intenctal EEG and PET scan. (e) Removal of the zone of cortical abnormality may result in control of medically refractory seizures, including generalized infantile spasms. REFERENCES Alvarez LA, Shinnar S,Mosht SL. Infantile spasms due to unilateral cerebral infarcts. Pediatrics 1987;79:1024-6. A u e m l d G, Schreck W, Dietrich HA. Vigabatrin in therapy-resistant childhood epilepsies. Epilepsia I99 I ;32(suppl 1 ):24. Bachman DS. Use of valproic acid in treatment of infantile spasms. Arch Neurol 1982;39:49-52. Blennow G, Stark L. High dose B6 treatment in infantile spasms. Neuropediatrics 1986;17:7-10. Branch CE, Dyken PR. Choroid plexus papilloma and infantile spasms. Ann Neurol 1979;5:302-4. Chiron C, Mondragon S, Dulac 0, Beaumont D, Palacios L, Luna D. Clinical trial of vigabratin in infantile spasms. Epilepsia 1989;30:661. Chugani HT, Shields WD, Shewmon DA, Olson DM, Phelps ME, Peacock WJ. Infantilespasms, I: PET identifiescortical dysgenesk in cryptogenic cases for surgical treatment. Ann Neurol1990;27: 406-13. Commission on Classificationand Terminology of the International League Against Epilepsy. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 1989;30389-99. Cusmai R, Dulac 0, Diebler C. Lesions focals dans les spasmes infantales. Neurophysiol Clin 1988;18:235-41. Dreifuss FE. Pediatric qileptobgy. Littleton:John Wright-PSG, 1983: 104-6. Dreifuss FE, Fanuell J, Holma G, et al. Infantile spasms: a comparative trial of nitrazepam and corticotropin. Arch Neurol 1986;43:1 107-10. Dyken PR, DuRant RH, Minen DB, King DW. Short term effects of valproate on infantile spasms. Pediatr Neurol 1985;1:34-7. Engel J, Jr. Seizures and epilepsy. Philadelphia: F. A. Davis, 1989: 202-3. Ferry PC,Banner W, Wolf RA. Seizure disorders in children. Philadelphia: J. B. Lippincon, 198637. Fois A, Malandrini F,Balcstri P, Giorgi D. Infantile spasms-longterm results of ACTH treatment. Eur J Pediar 1984;142:51-5. Glaze DG, Hrachovy RA, Frost JD, Kellaway P, Zion TE. Prospective study of outcome of infants with infantile spasms treated during controlled studies of ACTH and prednisone. J Pediatr 1988;I 12: 389-96. Holmes GL. Diagnosis and management of seizures in children. Philadelphia: W.B. Saunders, 1987:220-2. Hrachovy RA, Frost JD. Infantile spasms. Pediatr CIin North Am 1989;36:311-29. Livingston HJ, h u m o n t D, Admanoglou A, Aicardi J. Vigabatrin in the treatment of epilepsy in children. Br J Clin Pharmacol I98%27: 109s- 12s.
TREATMENT OF INFANTILE SPASMS Lombroso CT. A prospective study of infantile spasms: clinical and therapeutic correlations. Epilepsia 1983:24:135-58. Matsumoto A, Watanabe K, Negoro T, et al. Long-term prognosis after infantile spasms: A statistical study of prognostic factors in 200 cases. Dev Med Child Neurol 198 1 ;23:51-65 Mimaki T, Ono J, Yabuuchi H. Temporal lobe astrocytoma with infantile spasms. Ann Neurol 1983;I4:695-6. Ohtahara S, Yamatogi Y, Ohtsuka Y, Oka E, Takashi 1. Prognosis of West syndrome with special reference to Lennox syndrome: a developmental study. In Wada JA, Penry JK, eds. Advances in epilep1ology. Xrh Epilepsy lnrernational Symposium. New Y ork Raven Press, 1980149-54. Ohtsuka Y, Matsuda M, ogino T, Kobayashi K, Ohtahara S.Treatment ofthe West syndrome with highdose pyridoxal phosphate. Bruin Dev 1987;9:418-21, Palm DG, Korinthenberg R. West syndrome and Lennox-Gastaut syndrome in children with porencephalic cysts: long term followup after neurosurgical treatment. In: Niedermeyer E, Degen R, eds. The Lmnox-Gastaut syndrome. New York: Alan R. Liss, 1988:4 19-26. PaNone L, Incorpora G, LaRosa M, LiVolti S,Mollica F. Treatment of infantile spasms with sodium dipropylacetic acid. Dev Med Child Neurol I98 I ;23:454-61. Riikonen R. A long-term follow-up study of 214 children with the syndrome of infantile spasms. Neuropediatrics 1982;13: 14-23.
S3 I
Riikonen R, Donner M. ACTH therapy in infantile spasms: side effects. Arch Dis Child 198055664-72. Seki T, Kawahara Y, Hirose M. T h e long-term prognosis of infantile spasms-the present condition of cases of infantile spasms followed in school age. Folia Psychiatr Neurol Jpn 1976;30:297306. Shewmon DA, Altman K, Olson DM, Shields WD. Selective carotid Amytal suppression of independent multifocal spikes in children. Epilepsia 1990;31:653. Shields WD, Shewmon DA, Chugani HT, et al. Surgical treatment of pediatric epilepsy at UCLA. Epilepsiu I990;3 1 :652. Shields WD, Shewmon DA, Chugani HT, Peacock WJ. The role of surgery in the treatment of infantile spasms. J Epilepsy 1992; 3(suppl):321-4. Siemes H, Spohr HL, Michael T. Nau H. Therapy of infantile spasms with valproate: results of a prospective study. Epilepsia 198829: 553-60. Singer WD, Haller JS, Sullivan LR, Wolpert S, Mills C, Rabe EF. The value of neuroradiology in infantile spasms. J Pediarr 1982;100:47-50. Snead OC Ill. Treatment of infantile spasms. Pediarr Neurol I990:6: 147-50. Snead OC 111, Benton JW, Myers GJ. ACTH and prednisone in childhood seizure disorders. Neurology 1983;33:966-70. West WJ. On a peculiar form of infantile convulsions. Lanrer 184 I ;I : 724-5.
Epilepsia. Vol 33. Suppl. 4, 1992
Treatment of Infantile Spasms: Medical or Surgical? *tW. Donald Shields, *?D. Alan Shewmon, *?$.Harry T. Chugani, and §Warwick J. Peacock Departmenls of +Neurology, f Pediatrics, $Radiology, and §Neurosurgery, University of CaliJornia, Los Angeles, School of Medicine, Los Angeles, California, U.S.A.
Summary: Although infantile spasms were initially described in 1 84 I , remarkably little progress has been made in understanding the pathophysiology of this “peculiar form of infantile convulsions.” Consequently, our ability to treat infantile spasms is limited. Infantile spasms are classified as a “generalized” seizure disorder in the international classification system, which suggests that the underlying brain abnormality causing the seizures also must be diffuse or generalized. As the classification suggests, there are many diffise, or multifocal, brain disorders related to infantile spasms, e.g, inborn errors of metabolism, hypoxic-ischemic brain injury, and developmental brain defects such as tuberous sclerosis or Aicardi’s syndrome. On the other hand, infantile spasms have been reported in which a localized brain abnormality was present, e.g., tumor, stroke, and trauma. On rare occa-
In 1841 a British physician, W. J. West, wrote the first description of infantile spasms in a letter to the editor of Lancet titled “On a Peculiar Form of Infantile Convulsions.” He described in a young child (his own son) the seizures that we now recognize as infantile spasms, including their most distressing associated feature-developmental decline (West, 184 1). It is surprising how little has been learned of the basic mechanisms of these seizures in the 150 years since the original description. Perhaps as a consequence of our lack of understanding, few effective medical therapies are available, and the majority of the children still suffer the grave developmental consequences Dr. West described. Epilepsy in some children can be described as “catastrophic” because of the impact it has on their lives. Catastrophic childhood epilepsy is characterized by (a) onset early in life (often by age 1 year); (b) inadequate
sions, removal of a tumor has resulted in cessation of the generalized infantile spasms. This finding suggests that focal cortical abnormalities can cause infantile spasms and that removing the abnormality can stop the seizures. At University of California, Los Angeles, the Pediatric Epilepsy Surgery Program has developed new approaches to the treatment of infantile spasms. The principal underlying concepts are (a) children with medically refractory infantile spasms may have an area of cortical defect (called the zone of cortical abnormality) that causes the seizures and (b) infantile spasms are usually generalized seizures. Thus, the goal of the surgical assessment is not the identification of the focus of seizure onset but rather the identification of the zone of cortical abnormality. Key Words: Infantile spasms-AnticonvulsantsNeurosurgery.
response to antiepileptic drugs (AEDs); and (c) a poor developmental outcome. The most important catastrophic childhood seizure disorder is that of infantile spasms. The seizures of infantile spasms consist of brief bilateral jerks of the trunk and limbs, either flexion, extension, or a combination of both. Hypsarrhythmia, an electroencephalographic (EEG) abnormality defined as generalized, high-voltage, chaotic slowing with very high-voltage multifocal epileptiform discharges, is usually present during the interictal state but its presence is not required for making the diagnosis. The widespread EEG abnormalities combined with the bilateral clinical appearance of the seizures has led to the classification of infantile spasms as a generalized seizure disorder (Commission, 1989). Although neither the physiologic mechanisms of the seizures nor the specific reasons for the developmental consequences are understood, much has been learned about the neurologic diseases associated with infantile spasms. It is important to recognize that infantile spasms are a manifstation of an underlying brain disorder rather than a disease entity unto itself. Thus, recognition that a child has infantile spasms is the beginning of the evaluation rather than the conclusion. That
Address correspondence and reprint requests to Dr.W.Donald Shields at Division of Neurology, Department of Pediatrics, UCLA School of Medicine. 10833 Le Cone Avenue, Los Angeles, CA 90024-1752, U.S.A.
S26
TREATMENT OF INFANTILE SPASMS
diverse brain abnormalities are assOciated with infantile spasms is evidence, at least in some cases, that it is a secondary generalized seizure disorder rather than a primary one. On the basis of the identified or presumed cause of the seizures, patients with infantile spasms are divided into two groups: those with symptomatic and those with cryptogenic seizures. In symptomatic cases, an underlying cause can be identified, and in cryptogenic cases, no predisposing abnormality can be discovered after complete evaluation. Virtually anything that can injure the developing brain has been described as a cause of infantile spasms. As one might expect, many of these causes are diffuse or multifocal (e.g., hypoxic brain damage, diffuse brain malformation, inborn error of metabolism), but others are unifocal brain disturbances (e.g., tumor, stroke, trauma, focal brain malformation). Regarding patients who have a unifocal brain disorder such as tumor or stroke, one can ask an important question: Is the seizure disorder a manifestation of an undetected generalized brain lesion or can a unifocal brain disorder result in a generalized seizure lesion such as infantile spasms? A corollary question logically follows: If the patient’s seizures are medically refractory, can resection of the unifocal brain lesion result in cessation of the infantile spasms? MEDICAL TREATMENT OF INFANTILE SPASMS
Infantile spasms are often difficult to control with medications (Bachman, 1982; Dreihss et al., 1986; Hrachovy and Frost, 1989).One factor setting infantile spasms apart from all other seizure disorders is the superior efficacy of steroid drugs rather than AEDs in treatment (Dreifuss, 1983;Ferry et al., 1986;Holmes, 1987;Engel, 1989).There is disagreement as to whether corticotropin is more effective than prednisone, but most authors recommend 20-40 units of corticotropin per day given intramuscularly as the initial treatment. Many patients respond dramatically to corticotropin, with cessation of seizures after one or two doses, but treatment failures are common. Medications reported to be occasionally effective include pyridoxal phosphate (Blennow and Stark, 1986;Ohtsuka et al., 1987),highdose corticotropin (>lo0U/day) (Riikonen and Donner, 1980;Snead et al., 1983;Snead, 1990), highdose valproate (PaNone et al., 1981; Dyken et al., 1985; Siemes et al., 1988), and nitrazepam, y-vinyl-yaminobutync acid (Chiron et al., 1989;Livingston et al., 1989;Auerswald et al., 1991).Infantile spasms also have been reported to stop spontaneously, so evaluation of a therapeutic regimen that controls only a small percentage of patients is difficult.
S2 7
PROGNOSIS
The prognosis for normal development in children who develop infantile spasms is poor. Overall, 90-95% of children with infantile spasms will be moderately or markedly mentally retarded by age 5 years (Glaze et al., 1988).The prognosis appears to be related primarily to the cause and to seizure control (Riikonen, 1982; Fois et al., 1984). Other factors, such as age of onset of seizures, the presence of seizures other than infantile spasms, and localized EEG, computed tomographic (CT), and magnetic resonance imaging (MRI) abnormalities have also been related to outcome. The most important predictor of outcome is the underlying cause. In about 60-70% of patients, the cause is categorized as symptomatic and in 30% as cryptogenic. The probability of normal outcome for children with symptomatic infantile spasms is only about 5%, whereas it is as high as 38% in those with cryptogenic seizures (Glaze et al., 1988). In either group, inability to control the seizures indicates a poorer prognosis for normal development. Onset of seizures before the child is 4-6 months of age is not a good sign. In one study, seven of nine patients with an onset before 6 months of age were severely handicapped at follow-up, whereas only one of 16 with onset at the age of 6 months or older had such severe handicaps (Seki et al., 1976). In another study, one of 78 children with onset before 4 months of age developed normally, compared with 30 of 195 children with seizure onset at 4 months or older (Lombroso, 1983). The Occurrence of other types of seizures in addition to infantile spasms is also unfavorable. Three studies found that if other types of seizures occurred before or after infantile spasms,15% or fewer patients had a good outcome, whereas if infantile spasms were the only seizure type,more than 30% of the patients had a good outcome (Ohtahara et al., 1980; Matsumoto et al., 1981; Riikonen, 1982). Thus, patients who have medically refractory symptomatic infantile spasms and onset of seizures before the age of 4-6 months have a poor prognosis for normal development. For this group of patients, new therapies are certainly needed. SURGICAL TREATMENT: BACKGROUND
Evidence of localized brain abnormalities has been reported in many patients with infantile spasms. Focal disturbances have been demonstrated on EEG, localized structural defects by CT and MRI,and focal metabolic abnormalities by positron emission tomography (PET).EEG focal disturbance is the most commonly reported abnormality. Such focal abnormalities were EpiIepsia, Vd.33, Supd. 4, I992
W. D. SHIELDS ET AL.
S28
noted in one series in 110 of 149 cases (Riikonen, 1982). Because the EEG of a child with infantile spasms is so abnormal, it is easy to miss a relatively subtle EEG abnormality such as focal slowing or decreased activity. Structural defects detected by CT or MRI are less common but usually much more obvious (Singer et al., 1982). Alvarez et al. (1987) reported three children with porencephaly who developed infantile spasms and proposed that the porencephaly was the cause of the infantile spasms. Cusmai et al. (1988) found CT evidence of focal abnormalities in 17 of 174 cases. In a retrospective review of 37 patients with infantile spasms, we observed definite localized abnormalities on CT or MRI in 17 cases (46%) and EEG evidence of focality in another 6 (16%). Only 38% of the patients had no localizing features (Shields et al., 1992). The percentage of localizing features we observed is probably an overestimate of the true incidence, as some of the patients were referred because they had evidence of a localizable lesion and were thus candidates for surgery. Nevertheless, localized abnormalities may be much more common than previously appreciated. There are a few case reports of children with infantile spasms in whom a localized brain lesion was observed on neuroimaging and whose spasms stopped when the lesion was resected. Infantile spasms ceased after surgical resection of a right temporal astrocytoma in one case (Mimaki et al., 1983), by resection of a choroid plexus papilloma in another (Branch and Dyken, 1979), and by fenestration of porencephalic cysts in four cases (Palm and Korinthenberg, 1988).
SURGICAL TREATMENT: CONCEP'IS The Pediatric Epilepsy Program at University of California, Los Angeles (UCLA) was developed for the special needs of children with seizures. Pediatric epilepsy surgery evolved as part of the program and as a complement to the adult epilepsy surgery program. New concepts and protocols have been developed for addressing the unique challenges presented by children with medically refractory seizures. During the past 30 years, surgery to control medically intractable epilepsy has been directed toward patients with partial seizures (which may generalize secondarily). The essential observation that permits surgery is the identification of the area of seizure onset by surface EEG or by the placement of electrodes in or on the brain. Patients with generalized seizures (either primary or secondarily) have therefore not qualified for f d resective surgery. One result of our program has been the recognition that some patients with generalized seizures such as infantile spasms have a localizable Epilepsia Vd.33, Supd. 4, 1992
cortical abnormality that can be resected, with consequent control of seizures.The ability to perform surgery in these patients has necessitated fundamental changes in the concepts used in identifying the area to be removed. Through the study of a series of patients, we have determined factors related to the underlying disorders that sometimes cause infantile spasms and have developed new protocols that identify the area of abnormal cortex responsible for the seizures. We have called this area the zone of cortical abnormality to distinguish it from the epileptic focus or epileptic zone used to identify surgical candidates among patients with partial seizures. At UCLA, cortical resection has become an accepted method of therapy for medically intractable infantile spasms. The two concepts that govern surgical therapy in children with infantile spasms can be summarized as follows. Concept 1: Because the patients may have generalized or multifocal bihemispheric seizures, the goal of presurgical assessment is not the identification of the focus of onset of seizures but rather the identification of an area of structurally abnormal cortex called the zone of cortical abnormality. Concept 2: Resection of the zone of cortical abnormality in properly selected patients can result in the cessation of generalized seizures, including infantile spasms.
SURGICAL CANDIDATES Children are considered for evaluation for surgical control of seizures if they have infantile spasms that are refractory to corticotropin and/or prednisone and to trials with AEDs. We require that no evidence exists of degenerative, metabolic, or storage diseases, or medical contraindications for surgery. The majority of the patients have evidence of a localized brain abnormality before undergoing formal surgical evaluation. Most patients had partial seizures that antedated the onset of infantile spasms and/or had focal abnormalities on EEGs performed before the development of hypsarrhythmia. We have evaluated several patients who did not have preexisting evidence of localized brain defect and have not been able to identify a zone of cortical abnormality in the majority of such patients.
SURGICAL EVALUATION Evidence of a zone of cortical abnormality is sought by careful neurologic examination, by review of the history of seizures for determining whether partial seizures preceded or followed the infantile spasms, and by examination of present and past EEGs for identifying f d abnormalities. CT and MRI scans also are examined for evidence of a focal structural lesion. If the quality of CT and MRI scans is inadequate, they
TREA TMENT OF INFANTILE SPASMS
are repeated. Children whose results of neuroimaging studies reveal diffuse damage affecting both cerebral hemispheres are not further considered for surgery. The children are then admitted for in-hospital, 24h/day closed-circuit television EEG (CCTVIEEG) to characterizeand quantify seizures. Because the children have very frequent seizures, the duration of CCTV/ EEG is usually much shorter than the typical evaluation for epilepsy surgery. While the child is hospitalized, a PET scan is performed to evaluate the metabolic activity of the brain, particularly identification of focal hypometabolism.
SURGICAL CRlTERIA Children are recommended for surgery if there is a confluence of data from their history and EEG, C W / EEG, CT, MRI, and PET studies indicating the presence of a zone of cortical abnormality (a single lobe, contiguous multilobes, or one hemisphere). No single abnormal result is sufficient for surgery. A surgical decision is based on the following criteria.
EEGs and preoperative CC”/EEG (a) If the child had partial seizures prior to infantile spasms, the area of onset is identified by reviewing the description of the seizures and the past EEGs. (b) If the child has partial seizures at the time of surgical evaluation, the CCTV/EEG is reviewed to identify the focus of onset of partial seizures. (c) If the child has generalized or multifocal seizures at the time of surgical evaluation, the CCTV/EEG is analyzed for the presence of interictal abnormalities associated with a structural defect (focal slowing, a predominant focus of interictal epileptiform discharges, decreased P activity, or unilateral subclinical electrodecremental events). Imaging studies (a) The CT and/or MRI scans are examined for evidence of a zone of cortical abnormality. The abnormalities may be quite subtle (e.g., thickened cortex, narrow white matter, or suIci oriented in a wrong direction). If a zone of cortical abnormality is observed, it must correlate with the EEG focus. However, a completely normal CT and/or MRI result is not a contraindication for surgery. (b) The PET study must demonstrate a focal or hemispheric hypometabolic area (or hypermetabolic if it is an ictal study rather than i n t e n d ) that correlates with the focal EEG abnormality and with the CT/MRI (if abnormal). If the PET is the only study indicating focality, the child is not considered to be a candidate for surgery.
S29
(c) The brain outside the area of possible resection should appear to be relatively normal on CT,MRI, and PET (excluding generalized atrophy due to corticotropin therapy). Neurologic examination If a focal neurologic deficit is observed on examination, it must correlate with the zone of cortical abnormality being considered for resection. Additional neuropbysiologic studies If the child is a likely surgical candidate but the above do not define the zone of cortical abnormality sufficiently, the following studies are performed: (a) somatosensory-evoked potentials, observing for unilateral decreased or absent cortical peak; (b) thiopental sodium test showing a focal area of impaired drug-induced B activity on EEG, (c) intracarotid amobarbital injection on the “bad side,” resulting in suppression of multifiocal independent spikes bilaterally (Shewmon et al., 1990). Assessment of anticipated neurologic deficit The cortical resection cannot create an unacceptable new neurologic deficit. There is, however, no absolute definition of “unacceptable.” A decision is made in consultation with the patient’s parents on the basis of the severity of the seizures, the extent of developmental failure, and the prognosis if the seizures are to remain uncontrolled, compared with the effects of the anticipated deficit. For example, children considered for hemispherectomy are required to manifest hemiparesis prior to surgery. Thus, creation of a new severe hemiparesis is not considered “acceptable.” However, creation of a homonymous hemianopsia is usually considered acceptable.
SURGICAL TREATMENT: RESULTS Since 1986, we have performed surgery on 20 patients who bad a history of infantile spasms. The first four patients had partial seizures at the time of the surgery but had a history of generalized infantile spasms (Shields et al., 1992). This finding suggested that some children with infantile spasms had a localized cortical abnormality associated with the generalized seizures. The generalized seizures were controlled with corticotropin, but catastrophic partial seizures persisted. We questioned whether there were children with medically refractory generalized infantile spasms who had a similar localized abnormality that could be removed, with consequent control of the generalized seizures. In 1987, a patient with tuberous sclerosis presented with medically refractory generalized infantile spasms. An MRI scan demonstrated the typical areas of signal intensity in the cortex, but there was also a cortical glioma in the left parietal lobe that required resection. Epilepsia Vd.33, Suppl. 4. 1992
S30
W.D. SHIELDS ET AL.
The EEG revealed that the area near the glioma had a very active interictal epileptogenic focus. We determined that if this patient’s infantile spasms resolved with resection of the cortical glioma, the essential concept of control of generalized seizures by resection of a localized cortical abnormality would be supported. The glioma was resected, and the patient was seizurefree for 7 months. Following a period of seizure recurrence that was associated with the discontinuation of nitrazepam therapy, he again became seizure-free and has been so for the last I $ years. We subsequently extended the surgery protocol to include patients with medically refractory generalized infantile spasms even without associated focal seizures. Of the next 16 children, 13 had generalized infantile spasms at the time of surgery. The other three had mixed partial seizures and infantile spasms. The results with respect to seizure control have been quite gratifying. Most patients have a large zone of cortical abnormality requiring multilobar resection or hemispherectomy (Shields et al., 1992). Of the initial nine previously reported patients who were treated surgically for medically intractable infantile spasms, seven remain free of seizures 3-6 years after surgery. The surgery is as effective in patients who have generalized seizures as in patients who have partial seizures at the time of surgery (Shields et al., 1990). Several patients had cryptogenic infantile spasms with normal findings on CT and MRI Scans and no medical history suggesting a cause for the infantile spasms. The four reported by Chugani et a!. (1 990) all became seizurefree after surgery. CONCLUSIONS
The obvious answer to the question posed by the title of this article is that both types of treatment, medical or surgical, could be appropriate, depending on the specifics of the patient. Treatment, however, occurs sequentially-first medical and then the question of possible surgery. Without question, the initial treatment of choice is medical. Administration of corticotropin and/or prednisone should begin immediately after the diagnosis of infantile spasms has been established. Children who do not respond to this initial therapy should be given AEDs. Those whose seizures are controlled with medications are not surgical candidates whether or not they have an identifiable zone of cortical abnormality. On the other hand, the prognosis for children for whom these medical therapies fail is quite dismal. Consequently, the risks of surgery become acceptable when medical therapies have failed. Most patients with medically intractable infantile spasms are probably not candidates for surgery. The Epilepsia. Vol. 33. Suppl. 4,1992
incidence of the presence of a zone of cortical abnormality in children with medically refractory infantile spasms has not been established. However, on the basis of our experience, many such patients will benefit from surgery. We offer the following conclusions. (a) Some children with medically refractory infantile spasms have a localized brain defect, which we have called the zone of cortical abnormality. (b) Children with a zone of cortical abnormality may have generalized seizures. (c) Children with a zone of cortical abnormality may have normal CT and MRI scans. (d) The zone of cortical abnormality is often best identified with use of data obtained by intenctal EEG and PET scan. (e) Removal of the zone of cortical abnormality may result in control of medically refractory seizures, including generalized infantile spasms. REFERENCES Alvarez LA, Shinnar S,Mosht SL. Infantile spasms due to unilateral cerebral infarcts. Pediatrics 1987;79:1024-6. A u e m l d G, Schreck W, Dietrich HA. Vigabatrin in therapy-resistant childhood epilepsies. Epilepsia I99 I ;32(suppl 1 ):24. Bachman DS. Use of valproic acid in treatment of infantile spasms. Arch Neurol 1982;39:49-52. Blennow G, Stark L. High dose B6 treatment in infantile spasms. Neuropediatrics 1986;17:7-10. Branch CE, Dyken PR. Choroid plexus papilloma and infantile spasms. Ann Neurol 1979;5:302-4. Chiron C, Mondragon S, Dulac 0, Beaumont D, Palacios L, Luna D. Clinical trial of vigabratin in infantile spasms. Epilepsia 1989;30:661. Chugani HT, Shields WD, Shewmon DA, Olson DM, Phelps ME, Peacock WJ. Infantilespasms, I: PET identifiescortical dysgenesk in cryptogenic cases for surgical treatment. Ann Neurol1990;27: 406-13. Commission on Classificationand Terminology of the International League Against Epilepsy. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 1989;30389-99. Cusmai R, Dulac 0, Diebler C. Lesions focals dans les spasmes infantales. Neurophysiol Clin 1988;18:235-41. Dreifuss FE. Pediatric qileptobgy. Littleton:John Wright-PSG, 1983: 104-6. Dreifuss FE, Fanuell J, Holma G, et al. Infantile spasms: a comparative trial of nitrazepam and corticotropin. Arch Neurol 1986;43:1 107-10. Dyken PR, DuRant RH, Minen DB, King DW. Short term effects of valproate on infantile spasms. Pediatr Neurol 1985;1:34-7. Engel J, Jr. Seizures and epilepsy. Philadelphia: F. A. Davis, 1989: 202-3. Ferry PC,Banner W, Wolf RA. Seizure disorders in children. Philadelphia: J. B. Lippincon, 198637. Fois A, Malandrini F,Balcstri P, Giorgi D. Infantile spasms-longterm results of ACTH treatment. Eur J Pediar 1984;142:51-5. Glaze DG, Hrachovy RA, Frost JD, Kellaway P, Zion TE. Prospective study of outcome of infants with infantile spasms treated during controlled studies of ACTH and prednisone. J Pediatr 1988;I 12: 389-96. Holmes GL. Diagnosis and management of seizures in children. Philadelphia: W.B. Saunders, 1987:220-2. Hrachovy RA, Frost JD. Infantile spasms. Pediatr CIin North Am 1989;36:311-29. Livingston HJ, h u m o n t D, Admanoglou A, Aicardi J. Vigabatrin in the treatment of epilepsy in children. Br J Clin Pharmacol I98%27: 109s- 12s.
TREATMENT OF INFANTILE SPASMS Lombroso CT. A prospective study of infantile spasms: clinical and therapeutic correlations. Epilepsia 1983:24:135-58. Matsumoto A, Watanabe K, Negoro T, et al. Long-term prognosis after infantile spasms: A statistical study of prognostic factors in 200 cases. Dev Med Child Neurol 198 1 ;23:51-65 Mimaki T, Ono J, Yabuuchi H. Temporal lobe astrocytoma with infantile spasms. Ann Neurol 1983;I4:695-6. Ohtahara S, Yamatogi Y, Ohtsuka Y, Oka E, Takashi 1. Prognosis of West syndrome with special reference to Lennox syndrome: a developmental study. In Wada JA, Penry JK, eds. Advances in epilep1ology. Xrh Epilepsy lnrernational Symposium. New Y ork Raven Press, 1980149-54. Ohtsuka Y, Matsuda M, ogino T, Kobayashi K, Ohtahara S.Treatment ofthe West syndrome with highdose pyridoxal phosphate. Bruin Dev 1987;9:418-21, Palm DG, Korinthenberg R. West syndrome and Lennox-Gastaut syndrome in children with porencephalic cysts: long term followup after neurosurgical treatment. In: Niedermeyer E, Degen R, eds. The Lmnox-Gastaut syndrome. New York: Alan R. Liss, 1988:4 19-26. PaNone L, Incorpora G, LaRosa M, LiVolti S,Mollica F. Treatment of infantile spasms with sodium dipropylacetic acid. Dev Med Child Neurol I98 I ;23:454-61. Riikonen R. A long-term follow-up study of 214 children with the syndrome of infantile spasms. Neuropediatrics 1982;13: 14-23.
S3 I
Riikonen R, Donner M. ACTH therapy in infantile spasms: side effects. Arch Dis Child 198055664-72. Seki T, Kawahara Y, Hirose M. T h e long-term prognosis of infantile spasms-the present condition of cases of infantile spasms followed in school age. Folia Psychiatr Neurol Jpn 1976;30:297306. Shewmon DA, Altman K, Olson DM, Shields WD. Selective carotid Amytal suppression of independent multifocal spikes in children. Epilepsia 1990;31:653. Shields WD, Shewmon DA, Chugani HT, et al. Surgical treatment of pediatric epilepsy at UCLA. Epilepsiu I990;3 1 :652. Shields WD, Shewmon DA, Chugani HT, Peacock WJ. The role of surgery in the treatment of infantile spasms. J Epilepsy 1992; 3(suppl):321-4. Siemes H, Spohr HL, Michael T. Nau H. Therapy of infantile spasms with valproate: results of a prospective study. Epilepsia 198829: 553-60. Singer WD, Haller JS, Sullivan LR, Wolpert S, Mills C, Rabe EF. The value of neuroradiology in infantile spasms. J Pediarr 1982;100:47-50. Snead OC Ill. Treatment of infantile spasms. Pediarr Neurol I990:6: 147-50. Snead OC 111, Benton JW, Myers GJ. ACTH and prednisone in childhood seizure disorders. Neurology 1983;33:966-70. West WJ. On a peculiar form of infantile convulsions. Lanrer 184 I ;I : 724-5.
Epilepsia. Vol 33. Suppl. 4, 1992
