Electrophysiological characterization of spontaneous and carbamazepine-induced epileptic negative myoclonus in benign childhood epilepsy with centro-temporal spikes.

Abstract

Epileptic negative myoclonus (ENM), a transient muscular atonic phenomenon time-locked to epileptiform EEG abnormalities, is often observed in children with benign childhood epilepsy with centro-temporal spikes (BECTS). In some, for unknown reasons, ENM can be worsened by carbamazepine (CBZ). We describe two children aged 11 and 15 years, in whom CBZ precipitated seizure worsening and ENM. We investigated the morphological and topographic features of the EEG abnormalities while on CBZ and after CBZ withdrawal and compared them with those from 9 children with classical BECTS. The aim of the study was to identify possible electrophysiological specificities in patients who eventually develop ENM during CBZ treatment. The characterization of EEG abnormalities, related (R) and unrelated to ENM (U), in patients with ENM and rolandic discharges (RD) and in matched controls with BECTS was performed based on polygraphic digital EEG recordings. Off-line time-domain analysis included correlation coefficient between EEG and EMG channels, quantitative analysis on ENM, and topographic analysis on spike-and-wave complexes. Z-score test and paired t test were used when appropriate for statistical analysis on R, U and RD. Recordings in both children with BECTS and ENM while on CBZ showed frequent R discharges (mean interval between R=19.89+/-9.4 s in patient 1; 2.16+/-1.2 s in patient 2). Withdrawal of CBZ produced abatement of R (no R recorded in patient 1; 5.69+/-7.1 s in patient 2) and reduction of the slow wave component of R (P<0.01). Morphology and topography of R and RD differed in field distribution, amplitude (P<0.01) and duration (P<0.01) of the slow wave component. RD and U did not show a significantly different morphology and field distribution. Our findings suggest that an increased cortical inhibition could be the electrophysiological correlate of CBZ-induced ENM. If confirmed on a larger series, the presence of spike-wave (rather than sharp waves) discharges in children with BECTS might be used as an electrophysiological predictor of an abnormal response to CBZ.