doi: 10.1053/ejpn.2001.0403 available online at http://www.idealibrary.com on European Journal of Paediatric Neurology 2001; 5: 41±47 ORIGINAL ARTICLE Primary neonatal thalamic haemorrhage and epilepsy with continuous spike-wave during sleep: a longitudinal follow-up of a possible significant relation JOSE PAULO MONTEIRO, ELIANE ROULET-PEREZ, VERONIQUE DAVIDOFF, THIERRY DEONNA Neuropaediatric Unit CHUV, Lausanne, Switzerland Epilepsy with continuous spike-waves during sleep was diagnosed in a child who suffered primary neonatal thalamic haemorrhage, and who was followed from birth to 17 years of age. Early cognitive development was normal. Acquired behavioural problems and cognitive stagnation could be directly related to the epilepsy and not to the initial lesion and posthaemorrhagic hydrocephalus. This case and long-term follow-up data on a few children who suffered primary neonatal thalamic haemorrhage suggest that epilepsy with continuous spike-waves during sleep can be a sequel. Disturbances of thalamocortical interactions could play a role in the still poorly understood syndrome of epilepsy with continuous spike-waves during sleep. Keywords: Thalamic haemorrhage. Newborn. CSWS. Cognitive disturbances in children. Epilepsy. Introduction It is now increasingly recognized that epilepsy with continuous spike-waves during sleep (CSWS) in children with acquired cognitive disturbances can be seen with various focal cortical pathologies of early onset.1±2 It is, however, a rare occurrence, considering the large number of partial lesional epilepsies of prenatal or perinatal origin. This suggests that special aetiologic factors play a role in the clinical±epileptic syndrome of epilepsy with CSWS. We report a child who suffered primary neonatal thalamic haemorrhage (PNTH), a recently described condition,3 and who developed epilepsy with CSWS (Fig. 1). The long-term observation of this child, together with the suspected role of the thalamus in epilepsy and literature on follow-up of PNTH, suggests that early thalamic lesions might play a role in the genesis of this particular epileptic syndrome. Case report This right-handed boy is the elder of two children. There was no family history of epilepsy or neurological disease. He was born at term after an uncomplicated pregnancy by vaginal vertex delivery. Birthweight was 2950 g with a length of 48cm and a head circumference of 33cm. The Apgar scores were 9/10/10. His postnatal course was uneventful until the 13th day of life, when he was admitted with the sudden onset of seizures, opisthotonos, horizontal nystagmus, and ocular sunset phenomenon. Lumbar puncture showed haemorrhagic cerebrospinal fluid. A cranial ultrasound disclosed Received 3.7.2000. Revised 16.10.2000. Accepted 7.11.2000. Correspondence: Prof. Thierry Deonna, Neuropaediatric Unit ± Paediatric Department CHUV, Rue du Bugnon, CH-1011 Lausanne, Switzerland e-mail: Thierry.Deonna#chuv.hospvd.ch 1090±3798/01/05/0041+7 $35.00 & 2001 European Paediatric Neurology Society 42 Original article: J P Monteiro et al. Fig. 1. Summary of clinical history, with the course of epilepsy, anti-epileptic drugs, EEG results and behavioural and cognitive changes, from neonatal period to 17 years of age: PB: phenobarbital. CBZ: carbamazepine. VPA: valproate. PHT: phenytoin. LTG: lamotrigine. CSWS: continuous spike-waves during sleep. LF: left focus. N: normal EEG. FSIQ: full scale IQ. VIQ: verbal IQ. PIQ: performance IQ a large left thalamic haemorrhage and blood in both lateral ventricles. A computed tomography (CT) scan confirmed these findings and showed no associated cortical lesion (Fig. 2a). Cerebral angiography was not done. An electroencephalogram (EEG) showed left-sided sharp waves. Seizures were controlled with phenobarbital. A progressive asymmetrical ventricular dilatation, more marked on the left side, was stabilized after repeated lumbar punctures. Early development was normal. The child walked at 13 months, spoke in sentences at 24 months, knew colours and asked questions at the age of 3 years. A cranial ultrasound at 12 months showed stabilized hydrocephalus and residual echogenicity of the left thalamus. An EEG at 16 months was normal and phenobarbital was stopped. At 312 years he had a partial right motor seizure of 45 minutes duration, during the daytime. Phenobarbital was recommenced. An EEG showed a leftsided temporoparietal epileptic focus (Fig. 3a) but he continued to have right motor seizures and complex partial seizures, neither frequent nor severe. Cognitive and behavioural difficulties appeared and progressively increased between the age of 4 and 11 years (see neuropsychological data) and for the first time at 5 years a repeat EEG showed CSWS. At the age of 612 years, he progressively developed morning headaches, vomiting, ataxia and papilloedema. Magnetic resonance imaging (MRI) showed progression of his hydrocephalus and a ventriculocardiac shunt was performed at the age of 7 years. Figure 1 summarizes the successively tried antiepileptic drugs, which were neither efficient nor tolerated. At the age of 11 years, lamotrigine was introduced in another hospital in the hope of improving cognitive function and EEG abnormalities, but he developed a Stevens±Johnson syndrome with acute hepatic insufficiency and chronic cholestasis. A liver biopsy was compatible with `vanishing bile duct syndrome'. At this time all anti-epileptic drugs were stopped and prednisone was introduced at low dose (0.5 mg/kg on alternating days) for treating the liver disease and was mantained until the age of 17 years. The child gradually recovered. From the age of 14, he started having isolated short partial complex seizures, (2±3 per year) with a gustatory aura and motor component on the right. He was in a special school until the age of 16 years and is now in elementary professional training. At 17 years his neurological examination was normal. In Original article: Primary neonatal thalamic haemorrhage and epilepsy with CSWS 43 Fig. 2. Neuroradiological evolution. (a) CT scan at admission (13 days of life), showing left thalamic haemorrhage, with intraventricular bleeding (note: the left side is on your left). (b,c) MRI at 17 years showing the atrophic aspect of the left thalamus, residual ventricular dilatation more marked on the left side. The left hippocampus and parahippocampal region appears smaller than the right. particular, there were no signs of hemiparesis, hemidystonia, sensory loss or hemianopia. A repeat MRI (Fig. 2b,c) showed an atrophic lesion of the left thalamus and a residual hydrocephalus predominantly on the left side. The left temporal horn was enlarged and the left hippocampus and mesiotemporal region appeared smaller and undifferentiated compared with the right, but without signs of sclerosis. Electroencephalographic data During this 17-year follow-up, 25 recordings were obtained (three awake EEGs, 20 nap EEGs and two whole-night recordings). In the neonatal period, after the left thalamic lesion and evolution to posthaemorragic hydrocephalus, focal activity in the left hemisphere was observed, which 44 Table 1: Original article: J P Monteiro et al. Outcome of primary neonatal thalamic haemorrhage (PNTHD) reported in the literature PNTH reported in the literature No. of patients Follow-up Developmental delay Hydrocephalus Epilepsy Trounce (1985)3 Adams (1988)8 Roland (1990)9 Govaert (1992)10 Campistol (1994)11 Incorpore (1999)12 nˆ4 nˆ2 n ˆ 12 nˆ3 nˆ5 nˆ2 5±20m 10±15m 18m 5±6m 5m±11a 12m±6a 0/4 0/2 10/12 1/3 4/5 0/2 3/4 0/2 7/12 0/3 2/5 1/2 0/4 0/2 8/12 not characterized 0/3 3/5 1 partial; 2 generalized 2/2 1 CSWS; 1 generalized n ˆ 28 Median 24 m 15/28 13/28 13/28 m: months. y: years. progressively disappeared from around 16 months. A left parietotemporal focus reappeared again in an EEG performed when the child was 3 years old, after an episode of focal right motor status epilepticus. CSWS were recognized for the first time at 5 years of age. Abundant and generalized spike-wave activity was already present in the awake state, becoming almost continuous during sleep, particularly in the left hemisphere, over the left parietotemporal area, with frequent generalization to the right hemisphere. No clinical seizure was observed during the EEG recordings. Between 5 and 10 years, this intense epileptic activity was unchanged by successive medications (phenobarbital, valproate, carbamazepine, phenytoin, clobazam). Ventricular shunting, performed at the age of 7 years, brought a transient disappearance of CSWS, which reappeared 2 months later. During the whole observation period the background activity was preserved, except for a discrete slowing over the left hemisphere. After 11 years, while on low dose corticosteroids, improvement of clinical seizures was observed and CSWS gradually disappeared. The last EEG, performed at 17 years of age showed a left centrotemporal slowing, with rare focal spikes present only in the first part of the sleep cycle. Neuropsychological data Neuropsychological evaluations were performed about once a year between the age of 312 and 13 years. Global results were within the lower range of normal limits between the age of 312 and 6 years, decreased insidiously between the age of 7 and 10 years, and finally remained in the mildly deficient range. Raw scores never dropped (true deterioration), but did not progress at the expected rate (stagnation). The child was always considered slow, with the weakest scores in graphomotor, visuospatial organization (complex figure of Rey) and verbal memory tasks. He had neither visual perceptual problems nor gestural apraxia nor specific language impairment. Behaviour was considered normal until the age of 4 years. At this time, attention deficit, hyperactivity, disinhibition, irritability and fatigue were first noticed and worsened progressively. In addition defiance, aggression and provocative attitudes were observed between 7 and 10 years. These behavioural disorders caused fluctuations in performances which certainly played a major role in the low results obtained in the cognitive tests during this period. Figure 1 shows that the onset of behavioural problems coincided with the onset of CSWS between 3 and 5 years, followed 2 years later by stagnation of cognitive performances. Improvement in behaviour after the age of 11 years corresponded also to gradual cessation of the intense epileptic activity and stabilization of IQ scores. Discussion Our patient suffered from epilepsy with CSWS related to PNTH. The clinical course with normal development in the first years seems here mainly related to this form of epilepsy with its typical delayed and prolonged cognitive stagnation, rather than to the thalamic haemorrhage itself or posthaemorrhagic hydrocephalus. The MRI at 17 years, in addition to the thalamic lesion and residual hydrocephalus, also showed abnormalities in the hippocampal and mesiotemporal region, which probably also suffered damage in the neonatal period. However, temporal lobe seizures of hippocampal origin usually do not present with epilepsy with CSWS. We thus postulate that the thalamic pathology in addition to a focal cortical lesion were Original article: Primary neonatal thalamic haemorrhage and epilepsy with CSWS 45 Fig. 3. (a) Sleep EEG (14 electrodes) at 312 years showing left temporoparietal focus. (b) Sleep EEG (14 electrodes) at 5 years showing CSWS, more marked in the left temporoposterior region. 46 in our case responsible for this particular epileptic syndrome. Epilepsy with CSWS has initially been reported in children without evidence of brain lesion, but it is now increasingly recognized that this particular age related epileptic syndrome can be observed with various focal cortical pathologies.4±5 It has also been observed in shunted hydrocephalus,6 but in these cases, the hydrocephalus is usually associated with other cerebral pathologies which are more likely the cause of CSWS than the hydrocephalus per se.7 Aggravation of CSWS by antiepileptic treatment has also to be considered, but is very unlikely in our case, given the unchanged EEG pattern during several years with various drugs. In this context PNTH is particularly interesting in view of the still disputed role of the thalamus in the pathogenesis of epilepsy with CSWS. A survey of recent literature revealed 28 cases of PNTH (Table 1).3,8±12 Even though the follow-up is very short in most cases, 13/28 of the reported children (47%) had epilepsy. The rarer cases followed for a longer period11±12 had hydrocephalus and/or developmental delay and/or epilepsy. One of the children with a left thalamic haemorrhage (without hydrocephalus), also developed CSWS at the age of 6 years,12 but no data on a possible associated cortical lesion were given. Recently, four patients with neonatal thalamic haemorrhage lesions and epilepsy with CSWS between 3 and 5 years were reported (Boel M, 1999, unpublished data). In a study of seven children with CSWS and acquired aphasia or other cognitive disturbances, Maquet et al.13 found focal cortical/multifocal metabolic abnormalities on PET-scan and noted that these cortical asymmetries were not associated with the usual corresponding asymmetries in the thalamic nuclei found in focal epilepsies. These findings suggest either that only cortical neurons but not corticothalamic neurons are involved in the cerebral dysfunction generated by CSWS or that cortico-thalamic connections are disturbed. In an experimental study of seizures induced by bicuculline injections in the cortex of cats with unilateral thalamectomy, Steriade14 found an EEG seizure pattern similar to that found in intact animals, but interestingly also a `distinctive characteristic of displaying long sequences of continuous spike-waves activity at 2±4 Hz, with a spectacular synchronization across all (ipsilateral) cortical leads'. This observation has striking analogies with our clinical case of CSWS and posthaemorrhagic destruction of the thalamus and introduces new arguments supporting the theory of involvement of the thalamus in the Original article: J P Monteiro et al. pathogenesis of CSWS. Disturbed corticothalamic interactions may transform the physiological brain oscillations at work during slow sleep into paroxysmal epileptic activity and affect reverberating circuits which seem necessary to short term plasticity processes like memory.15 The case reported here and recent data12 ( Boel M, 1999, unpublished data) suggest that an early thalamic lesion can predispose to epilepsy with CSWS, probably in addition to cortical pathology and perhaps predisposing genetic factors. This hypothesis could be verified with longer followups of children with PNTH or with other types of pre- or perinatal thalamic lesions, which may not always be symptomatic or recognized in the neonatal period.16 Acknowledgements Neurological Service ± Centre Hospitalier Universitaire Vaudois 1011 Lausanne Neuroradiological Service ± Centre Hospitalier Universitaire Vaudois 1011 Lausanne. This work was supported by the Calouste Gulbenkian Foundation (JPM) and the Swiss National Research Fund ± Grants No 3227865.89 (TD) and 32-5299.97 (ER). References 1 2 3 4 5 6 Tassinari CA, Bureau M, Dravet C et al. Epilepsy with continuous spike and waves during slow sleep, otherwise described as ESES (epilepsy with electrical status epilepticus during slow sleep). In: Roger J, Bureau M, Dravet C et al. (eds) Epileptic Syndromes in Infancy, Childhood and Adolescence. London: John Libbey, 1992: 245±256. Jayakar PB, Seshia SS. Electrical status epilepticus during slow-wave sleep: a review. J Clin Neurophysiol 1991; 8: 299±311. Trounce JQ, Dodd KL, Fawer CL et al. Primary thalamic haemorrhage in the newborn: a new clinical entity. Lancet 1985; 1: 190±192. Deonna T, Davidoff V, Maeder-Ingvar M et al. The spectrum of acquired cognitive disturbances in children with epilepsy and continuous spike-waves during sleep. Eur J Ped Neurol 1997; 1: 19±29. Freitag C, Bast T, Ebinger F et al. Early acquired lesion brain lesion and continuous spike and wave activity during slow-wave sleep (CSWS): Clinical courses. Abstracts of the 26th Annual Meeting of the Society for Neuropediatrics. May 4±7 2000. Kiel, Germany. Neuropediatrics 2000; 31: A5. Veggioti P, Beccaria F, Papalia G et al. Continuous spikes and waves during sleep in children with shunted hydrocephalus. Child's Nerv Syst 1998; 14: 188±194. Original article: Primary neonatal thalamic haemorrhage and epilepsy with CSWS 7 Klepper J, BuÈsse M, Straûburg HM, Sorensen N. Epilepsy in shunt-treated hydrocephalus. Dev Med Child Neurol 1998; 40: 731±736. 8 Adams C, Hochhauser L, Logan WJ. Primary thalamic and caudate hemorrhage in term neonates presenting with seizures. Pediatr Neurol 1988; 4: 175± 177. 9 Roland EH, Flodmark O, Hill A. Thalamic hemorrhage with intraventricular hemorrhage in the fullterm newborn. Pediatrics 1990; 85: 737±742. 10 Govaert P, Achten E, Vanhaesebrouck P et al. Deep cerebral venous thrombosis in thalamo-ventricular hemorrhage of the term newborn. Pediatr Radiol 1992; 22: 123±127. 11 Campistol J, Pires M, PoÂo P, Iriondo M. Hemorragia talaÂmica e intraventricular en el recieÂn nacido a teÂrmino. Rev Neurol (Barcelona) 1994; 22: 673±675. 12 Incorpore G, Pavone P, Smilari PG et al. Late primary unilateral thalamic hemorrhage in infancy: 47 Report of two cases. Neuropediatrics 1999; 30: 264± 267. 13 Maquet P, Hirsch E, Metz-Lutz MN et al. Regional cerebral glucose metabolim in chidren with deterioration of one or more cognitive functions and continuous spike-and-wave discharges during sleep. Brain 1995; 118: 1497±1520. 14 Steriade M, Contreras D. Spike-wave complexes and fast components of cortically generated seizures. I. Role of neocortex and thalamus. J Neurophysiol 1998; 80: 1439±1455. 15 Steriade M. Coherent oscillations and short-term plasticity in corticothalamic networks. Trends Neurosci 1999; 22: 337±345. 16 Van Wezel-Meijler G, Hummel TZ, Oosting J et al. Unilateral thalamic lesions in premature infants: risk factors and short-term prognosis. Neuropediatrics 1999; 30: 300±306.