Clinical staging and electroencephalographic evolution of continuous spikes and waves during sleep

Abstract

Currently, in continuous spikes and waves during sleep (CSWS) there is a lack of systematic assessments of the clinically relevant stages and the evolution of the electroencephalographic features. The aim of this study is to describe the evolution over time of clinical and electroencephalographic features in CSWS. We enrolled patients from our video-electroencephalography (EEG) monitoring unit with CSWS and with overnight EEG studies with at least one overnight assessment per year over a minimum period of 3 years. We studied clinical presentation and electroencephalographic features. We calculated the (1) spike-wave percentage (SWP) as the percentage of 1-s bins containing at least one spike-wave complex and (2) spike frequency (SF) as the number of spikes per 100 s. Nine children (six boys) met the inclusion criteria during a 15-year period. Seven (78%) had an abnormal development prior to the epilepsy onset, and in two (22%) seizures were the only presenting symptom. Median age at epilepsy onset was 2 years (range 2 days to 4 years), at neuropsychological regression 5.1 years (4-7.7 years), and at seizure freedom 8.6 years (6.5-11.4 years). Median duration and range of clinically relevant stages were as follows: dormant stage (birth-epilepsy onset median 2 years, range 2 days-4 years), prodromal stage (epilepsy onset-neuropsychological regression 3.9 years, range 0.9-7.7 years), acute stage (neuropsychological regression-seizure freedom 2.9 years, range 2.1-6.6 years), and residual stage (after seizure freedom). Seven patients (78%) had a structural lesion on neuroimaging. At last follow-up (median 11.4 years, range 7.2-20.3 years), eight patients (89%) were receiving antiepileptic treatment, and all patients had residual neurocognitive deficits. During the acute stage, SWP was 85%. Evolution of electroencephalographic patterns included increasing-decreasing, continuously elevated, and fluctuating patterns (33.3% each). There was good correlation between SWP and SF (Spearman correlation-coefficient = 0.942; p < 0.0001). SF, which can exceed 100%, reflected changes in electroencephalography pattern in more detail than SWP, which cannot exceed 100% and therefore has a ceiling effect. Our series systematically studied the age of occurrence of the significant clinical events. These may assist in defining clinical stages, which can provide a useful framework for future clinical trials in patients with CSWS. The severity of the epileptiform discharges on EEG did not always correlate with seizure frequency and severity; epileptiform discharges could be prominent after seizure freedom and fluctuated along the course of the disease. The values of SWP and SF correlated well, but SWP based on 1-s bins has the potential disadvantage of a ceiling effect.