Abstracts

Rationale: Little is known about the stability of spike counts in children with epilepsy. Before interpreting assessments of spike frequency in patients with interictal epileptiform activity, it is important to know whether this is a stable or variable measure. Knowledge of the natural variations in spike frequency may bear on interpretation of maneuvers such as attempted pharmacologic suppression of spikes and help avoid mistaking natural variation for a drug suppression effect. We investigated the night-to-night variation in spike frequency by comparing the spike counts during sleep in children undergoing 48-hour ambulatory EEG recording, comparing spike frequency on the first night to that of the second night. Methods: We analyzed twelve 48-hour ambulatory EEGs performed at Children s Hospital Boston between August 2009 and June 2010. Studies without spikes, studies during which seizures occurred, or studies that contained artifact that precluded accurate spike counting were excluded. No child s antiepileptic drugs were adjusted during the recordings. When distinct spike foci existed in the same child, these were counted separately. In order to compare comparable sleep segments, we visually counted all spikes occurring during the first night of the recording in the first 20 minutes after the occurrence of the first sleep spindle that appeared during evening/nighttime sleep. The first night s count was compared to the spike count obtained during the second night s sleep during a 20-minute segment defined by the same method. Results: 17 separate spike foci were counted in 12 children (age range: 3-19 years; mean: 8 years). The firing rate for 7 spike foci increased from the first night to the second night, decreased for 9 foci, and did not change for one focus. The mean absolute percentage change between nights was 23.7% (S.D. = 14.8, range: 0-60%). The coefficient of variation (standard deviation/mean) was 0.62 suggesting moderate variability. The intraclass correlation coefficient was 0.51, also suggesting significant variability (the value would be 1 if measurements were completely stable between nights). Conclusions: In our group of 12 children with 17 spike foci, significant variability in spike frequency was seen during comparable segments of stage II sleep, despite the lack of intervention (e.g., medication changes) between the two nights. The mean percentage change of 23.7% and high coefficient of variation and intraclass correlation of these measures suggest significant natural variation in this phenomenon. Conclusions regarding observed changes in spike counts in the clinical setting should be made with caution.