EEG is a useful predictor of seizure recurrence following withdrawal of anticonvulsant medication in children

Abstract

Neurologists, like other mortals, are not clairvoyant. Attempting to predict future clinical events is nonetheless an inescapable reality of medical practice. So too is basing current management decisions on such predictions. A salient example is whether to discontinue anticonvulsant medication in an epileptic child whose seizures have been controlled for a prolonged period. In weighing the risk of seizure recurrence that attends drug withdrawal against the likelihood of neurologic and other systemic toxicity with drug continuation, a variety of factors may need to be considered. These frequently include seizure type, etiology, age of onset, and duration of epilepsy prior to treatment (1). Whether EEG findings assist in this decision-making process remains controversial, with some investigators (2-6) but not others (7-10) contending that it is a useful predictor of seizure recurrence. Comparing these studies is difficult because of the varying methodologies employed by the investigators. For example, the subjects may be selected from a variety of sources: epilepsy referral centers, neurology clinics, and less often from general medical practices. Also, although almost all of the major studies rely on EEGs obtained using standard recording techniques, the number and timing of recordings differ. For example, whereas data are commonly derived from EEGs taken just prior to drug withdrawal, some authors have gone further by comparing these results with those obtained prior to and following the institution of drug treatment. Still others continue periodic EEG monitoring during and following completion of the weaning process. In addition, methods of scoring EEG results range in complexity from the