We determined the additional yield of ambulatory over routine electroencephalography recordings in predicting seizure recurrence after antiepileptic drug (AED) withdrawal in 15 adult patients with various epilepsy syndromes who had been seizure free for at least 3 years (median=10 years). Eleven of 15 patients (74%) relapsed during or after AED withdrawal. All six patients with epileptiform discharges on ambulatory electroencephalography prior to AED withdrawal relapsed, compared with five of nine patients without epileptiform discharges. Ambulatory electroencephalography significantly increases the yield in detecting epileptiform discharges (n=6) compared with routine electroencephalography (n=1). A negative electroencephalographic finding is only a poor predictor of seizure freedom following AED withdrawal. On clinical grounds, our data suggest that patients with epilepsy, learning disability, and other known risk factors (history of abnormal EEGs and frequent seizures, abnormal MRI) are at great risk of seizure relapse during AED taper, irrespective of very long periods of seizure freedom.