Objective: To explore the relationship of seizure occurrence with candidate seizure precipitants in a prospective diary study, and to determine the relationship of precipitants to seizure self-prediction.
Methods: Eligible subjects were 18 or older, had localization-related epilepsy, at least one seizure within 12 months, and were able to maintain a daily diary. Information collected included the occurrence, time and characteristics of all seizures, hours of sleep, medication compliance, stress, anxiety, alcohol use, menstruation, and seizure self-prediction. Each night, subjects reported their estimate of the likelihood of a seizure the next day (self-prediction). Logit-normal models with a random subject-specific intercept were used to estimate an OR for the association of precipitants with seizure occurrence.
Results: Seventy-one subjects returned 15,179 complete diary days. For each hour of increased sleep on the preceding night, the relative odds of a seizure the following day decreased (OR 0.91, 95% CI 0.82, 0.99). One-unit increments of stress and anxiety (on a 10-point scale) were associated with an increased risk of seizure the following day (OR 1.06, 95% CI 1.01, 1.12 and OR 1.07; 95% CI 1.02, 1.12). With self-prediction included in the model, self-prediction (OR 3.7; 95% CI 1.8, 7.2) and hours of sleep for the night prior to the seizure (OR 0.90; 95% CI 0.82, 0.99) remained significant.
Conclusion: Lack of sleep and higher self-reported stress and anxiety levels were associated with seizure occurrence. In a model that included self-prediction, less sleep, and self-prediction had significant effects, whereas stress and anxiety did not. The psychological and biologic mechanisms which link stress and anxiety to self-prediction of seizures requires further exploration. Ultimately, seizure prediction based on precipitants, premonitory features, and self-prediction may provide a foundation for preemptive treatment.