Diagnosis and treatment of nonconvulsive status epilepticus.

Nonconvulsive status epilepticus (SE) is not uncommon and comprises at least one-third of all cases of SE. However, nonconvulsive SE consists of very different syndromes, a common feature being the difficulty in making the diagnosis. In this review, nonconvulsive SE is divided into typical absence SE, complex partial SE, nonconvulsive SE in patients with learning difficulties (including electrical SE during sleep, atypical absence SE and tonic SE), and nonconvulsive SE in coma. These conditions have different prognoses and treatments. The diagnosis of these conditions is critically dependent on EEG. When the EEG demonstrates typical ictal patterns, the diagnosis is usually straightforward. However, in many circumstances the EEG has to be differentiated from encephalopathic patterns, and this differentiation can prove troublesome, although the clinical and electrographic response to treatment can prove helpful. Nonconvulsive SE in patients with learning difficulties possibly provides the greatest diagnostic difficulty; the clinical presentation can be subtle resulting in the diagnosis being frequently missed. Whether the neuronal damage that occurs in convulsive SE and in animal models of limbic SE also occurs in nonconvulsive SE in humans is still a matter of debate. There are critical differences between the animal models and the human condition. Indeed, the prognosis of nonconvulsive SE is usually dependent on the underlying aetiology rather than the persistence of electrographic discharges. Because of these doubts, a more conservative approach to the treatment of particular types of nonconvulsive SE (those with a better prognosis) has been taken in this article. Thus, in most instances, oral benzodiazepines for the treatment of typical absence SE and complex partial SE are recommended. In some circumstances intravenous medication is necessary, but in neither condition is anaesthetic coma recommended. This contrasts with nonconvulsive SE in coma in which a more aggressive approach is suggested. Until there are more relevant animal models, and controlled trials of conservative versus more aggressive treatment, treatment regimens for nonconvulsive SE will remain largely speculative.