Key points are not available for this paper at this time.
Abstract Objective Seizure clusters require prompt medical treatment to minimize possible progression to status epilepticus, increased health care use, and disruptions to daily life. Isolated seizures may exhibit cyclical patterns, including circadian and longer rhythms. However, little is known about the cyclical patterns in seizure clusters. This post hoc analysis of data from a long‐term, phase 3, open‐label, repeat‐dose safety study of diazepam nasal spray modeled the periodicity of treated seizure clusters. Methods Mixed‐effects cosinor analysis evaluated circadian rhythmicity, and single component cosinors using 12 and 24 h were used to calculate cosinor parameters (e.g., midline statistic of rhythm, wave ampitude, and acrophase peak). Analysis was completed for the full cohort and a consistent cohort of participants with two or more seizure clusters in each of four, 3‐month periods. The influence of epilepsy type on cosinor parameters was also analyzed. Results Seizure‐cluster events plotted across 24 h showed a bimodal distribution with acrophases (peaks) at ~06:30 and ~18:30. A 12‐h plot showed a single peak at ~06:30. Cosinor analyses of the full and consistent cohort aligned, with acrophases for both models predicting peak seizure activity at ~23:30 on a 24‐h scale and ~07:30 on a 12‐h scale. The consistent cohort was associated with increases in baseline and peak seizure‐cluster activity. Analysis by epilepsy type identified distinct trends. Seizure clusters in the focal epilepsy group peaked in the evening (acrophase 19:19), whereas events in the generalized epilepsy group peaked in the morning (acrophase 04:46). Together they compose the bimodal clustering observed over 24 h. Significance This analysis of seizure clusters treated with diazepam nasal spray demonstrated that seizure clusters occur cyclically in 12‐ and 24‐h time frames similar to that reported with isolated seizures. Further elucidation of these patterns may provide important information for patient care, ranging from improved patient‐centered outcomes to seizure‐cluster prediction.
Fountain et al. (Fri,) studied this question.