Can patients with epilepsy predict their seizures?

Objective: To examine seizure predictability in a cohort of adult patients using a prospective seizure diary study, to assess the validity of a patient's predictions, and to determine if a subgroup of patients with epilepsy were able to reliably predict their seizures. Methods: Eligible subjects were 18 or older, had localization-related epilepsy, had >= 1 seizure within 12 months, and were able to self-maintain a seizure diary. Seizure self-prediction was assessed by a rating scale: "Do you think you will have a seizure in the next 24 hours?" Seizure prediction and seizure occurrence were modeled as binary outcomes. OR, positive predictive value, and negative predictive value were calculated for group and individuals, adjusted for the within person correlations. Results: Seventy-one subjects returned 15,635 diary entries. A positive prediction of seizures was associated with a twofold increased risk of seizure (Cochran-Mantel-Haenszel OR 2.25 (1.91 to 2.65). Overall, the specificity of positive prediction was 83.2% while the sensitivity was 31.9%. Twelve subjects (21% of subjects who had seizures) demonstrated significant within person OR; Cochran-Mantel-Haenszel OR for positive prediction in this predictor subgroup 3.14 (2.53 to 3.89). Predictors were younger (p = 0.026) and had a higher seizure rate (p = 0.003) than nonpredictors. Conclusions: A significant subgroup of our adult patients with epilepsy were able to self-predict their seizures, which may yield novel therapeutic opportunities.