The estimates in the literature of the risk of seizure relapse after antiepileptic medications are withdrawn range from less than 10% to nearly 70%. There is also little coherence regarding predictors of successful medication withdrawal. We performed a meta-analysis of the published literature to date to determine the risk of relapse at 1 and 2 years after discontinuation of medications and to examine the strength of association between the risk of relapse and three commonly assessed clinical factors: age of onset of epilepsy, presence of an underlying neurologic condition, and an abnormal EEG. We established criteria for inclusion of a study in the analysis, and 25 studies met these criteria. Overall, the risk of relapse at 1 year was 0.25 (95% CI, 0.21 to 0.30) and at 2 years it was 0.29 (95% CI, 0.24 to 0.34). Relative to epilepsy of childhood onset, epilepsy of adolescent onset was associated with a relative risk of relapse of 1.79 (95% CI, 1.46 to 2.19). Compared with childhood-onset epilepsy, adult-onset epilepsy was associated with a relative risk of 1.34 (95% CI, 1.00 to 1.81). Patients with remote symptomatic seizures were more likely to relapse than patients with idiopathic seizures; the relative risk was 1.55 (95% CI, 1.21 to 1.98). An abnormal EEG was associated with a relative risk of 1.45 (95% CI, 1.18 to 1.79). Although these figures help provide an estimate of an individual's likelihood of relapse, they should not be used as the sole basis on which to make the decision on discontinuation of medications.(ABSTRACT TRUNCATED AT 250 WORDS)