Objective: The temporal distribution of seizures in women with localization-related epilepsy occurs periodically according to a model "clock" with the peak phase of occurrence corresponding to menstrual onset. The location and laterality of the epileptic lesion as well as patient age may affect periodicity.
Methods: Baseline data from seizure and menstrual diaries of approximately 3 months duration were obtained from 100 women enrolled in a trial of hormonal therapy for localization-related epilepsy. Durations of individual cycles were normalized to a common menstrual phase and period. Normalized data were then combined to create distributions evaluated by localization (lobar: temporal [TL], extratemporal [XL], multifocal [MF], unknown), lateralization (left, right, bilateral, unknown), and age. Distributions were evaluated with analysis of variance (ANOVA) and curve-fitted by nonlinear least squares cosinor analysis.
Results: A total of 71 patients had TL (left = 25, right = 29, bilateral = 17), 10 XL, 14 MF, and 5 unknown seizure foci. XL and MF seizures occurred randomly across the 28-day cycle. TL seizures (left = 875, right = 706) occurred nonrandomly (ANOVA p = 0.0003) and cyclically with peak occurrence near onset of menses ([value +/- SD] peak = 1.6 +/- 2.3 days, period = 27.0 days). Left-side TL seizures peaked cyclically at onset of menses (ANOVA p = 0.04, peak = 0.0 +/- 3.0 days, period = 30 days); right-side TL seizures occurred randomly. Age did not have a cyclical effect. Women below the median age had a significantly higher seizure rate than those above the median age.
Conclusion: Circalunar rhythms of seizures in women, and therefore, possibly strategies of hormonal treatments of catamenial epilepsy, vary with the neuroanatomic substrate of the seizure focus.