Neonatal convulsions pose a unique challenge to the treating physician, both at the diagnostic and therapeutic level. Approximately 3% of the children admitted to a neonatal intensive care setting are diagnosed with convulsions (Nunes et al, Arq Neuropsiquiatr 2008). This frequency is probably an underestimation because subtle seizures might remain undiagnosed. By far (50%), convulsions are seen in the global setting of hypoxic–ischaemic encephalopathy, but in a substantial proportion of the children, aetiology in the neonatal phase is not found. In those infants, a systematic search for metabolic and genetic causes should be started as soon as possible. In particular, metabolic causes should be ruled out, because early treatment is sometimes the only way to stop cognitive deterioration. Also, it becomes clear that channelopathies can present in early life with convulsions (Lagae, Eur J Ped 2008).
From the diagnostic point of view, many advances have been made with bedside amplitude-integrated EEG systems. Although these systems are very useful in the detection of seizures, they do not allow for a precise classification of the seizures. A combination of amplitude-integrated systems with conventional EEG remains the ultimate standard for epileptologists (see Shellhaas et al, J Paediatr 2008).
Treatment is another hot debate. It is widely recognised that both phenobarbital and phenytoin could still be considered as first line drugs. Lidocaine and benzodiazepines (midazolam) are often used by neonatologists, whereas the “classic” anti-epileptic drugs (valproate, topiramate and levetiracetam) are advocated more by paediatric neurologists (Bassan H et al, Paed Neurol 2008).