Epilepsy in hemiplegic cerebral palsy due to perinatal arterial ischaemic stroke

Abstract

Study of the cerebral palsies requires their separation into pathological entities at the least and ideally into clinico-pathological causal pathways (which has now become possible with modern imaging).1 The study of epilepsy in hemiplegic cerebral palsy by Wanigasinghe et al.2 appropriately selects from a population-based sample the group with early arterial stroke. It is generally accepted that many of these events occur in the immediate prenatal period for uncertain reasons but a proposed mechanism is placentally shed thrombi passing through the right to left shunt.3 The characteristic imaging finding is of a large full thickness cortical deficit often amounting to porencephaly which provides the substrate for epileptic seizures. In the absence of a clear risk factor (e.g. amnionitis) it is not known as to whether the severity of involvement may vary with the quality of antenatal care from one obstetric centre to another. It is interesting that middle cerebral arterial infarcts at a later age have less cortical involvement and cause less epilepsy. From many studies of early onset brain syndromes it is clear that cerebral cortical damage is commonly associated with three groups of impairments: epileptic, cognitive, and psychiatric. Their coexistence is so common that the presence of, for example, cognitive impairment tends to predict the presence of epilepsy. Similarly the presence of epilepsy generally predicts a lower IQ in congenital hemiplegia generally.4 Thus despite the positive features of this study it would have been helpful to have cognitive data and test its value as a predictor of seizure occurrence and remission in this specific group. In this study the finding of severity of motor deficit as a risk factor for epilepsy relates presumably to the extent of the infarct. Neonatal seizures are more a measure of the precise timing of the infarct than its extent and might well not correlate as was found in the study. The seizure semiology is of interest. It would be intriguing to correlate the high rate of spasms (56%) with cognitive outcome and whether electroencephalogram abnormality was unilateral. Not surprisingly, partial seizures are common but some of the specific features are important. Nausea and vomiting are generally regarded as uncommon features of seizures and are associated with involvement of the temporal neocortex. They were present in 50% of seizures in this series but in under 3% of a consecutive series of 178 patients with intractable temporal lobe epilepsy.5 Visual hallucinations indicate an occipital source of seizures. The observation that seven patients had aphasia is also of interest since it might imply that some aspects of language localization remain in the damaged hemisphere, but we need more detail of the timing of the speech arrest within the seizure to be sure of the significance of these observations. Only two patients had reflex focal seizures which are thought to be a common feature of seizures with congenital hemiplegia but needs careful enquiry to identify them. The close resemblance of one patient’s epilepsy syndrome to benign myoclonic epilepsy of infancy may, with other patients, allow insight into this rather problematic condition. It is helpful to know that just over half the study group had epilepsy with a rate of about 25% in all with congenital hemiplegia, but that many went into seizure remission leaving only 15% with active epilepsy after 10 years. The ability to predict such remission would be helpful in the context of surgical decision-making. Only two patients have had epilepsy surgery and one more is being considered for this, which in the context of an active paediatric epilepsy surgery programme (and both surgeries being performed under the age of 15mo) suggests that the requirement for surgery is quite low. This study therefore is part of a process whereby the broad category of congenital hemiplegia is split into its pathological components, thus allowing much more precise delineation of the phenotype, management, and prognosis.6 It also shows that careful clinical study of seizures can provide interesting evidence about semiology and localization.